1.Condensed Movies: Story Based Retrieval with Contextual Embeddings ⬇️
Our objective in this work is the long range understanding of the narrative structure of movies. Instead of considering the entire movie, we propose to learn from the key scenes of the movie, providing a condensed look at the full storyline. To this end, we make the following four contributions: (i) We create the Condensed Movie Dataset (CMD) consisting of the key scenes from over 3K movies: each key scene is accompanied by a high level semantic description of the scene, character face tracks, and metadata about the movie. Our dataset is scalable, obtained automatically from YouTube, and is freely available for anybody to download and use. It is also an order of magnitude larger than existing movie datasets in the number of movies; (ii) We introduce a new story-based text-to-video retrieval task on this dataset that requires a high level understanding of the plotline; (iii) We provide a deep network baseline for this task on our dataset, combining character, speech and visual cues into a single video embedding; and finally (iv) We demonstrate how the addition of context (both past and future) improves retrieval performance.
2.Hyperspectral Image Restoration via Global Total Variation Regularized Local nonconvex Low-Rank matrix Approximation ⬇️
Several bandwise total variation (TV) regularized low-rank (LR)-based models have been proposed to remove mixed noise in hyperspectral images (HSIs). Conventionally, the rank of LR matrix is approximated using nuclear norm (NN). The NN is defined by adding all singular values together, which is essentially a
$L_1$ -norm of the singular values. It results in non-negligible approximation errors and thus the resulting matrix estimator can be significantly biased. Moreover, these bandwise TV-based methods exploit the spatial information in a separate manner. To cope with these problems, we propose a spatial-spectral TV (SSTV) regularized non-convex local LR matrix approximation (NonLLRTV) method to remove mixed noise in HSIs. From one aspect, local LR of HSIs is formulated using a non-convex$L_{\gamma}$ -norm, which provides a closer approximation to the matrix rank than the traditional NN. From another aspect, HSIs are assumed to be piecewisely smooth in the global spatial domain. The TV regularization is effective in preserving the smoothness and removing Gaussian noise. These facts inspire the integration of the NonLLR with TV regularization. To address the limitations of bandwise TV, we use the SSTV regularization to simultaneously consider global spatial structure and spectral correlation of neighboring bands. Experiment results indicate that the use of local non-convex penalty and global SSTV can boost the preserving of spatial piecewise smoothness and overall structural information.
3.Data-Free Network Quantization With Adversarial Knowledge Distillation ⬇️
Network quantization is an essential procedure in deep learning for development of efficient fixed-point inference models on mobile or edge platforms. However, as datasets grow larger and privacy regulations become stricter, data sharing for model compression gets more difficult and restricted. In this paper, we consider data-free network quantization with synthetic data. The synthetic data are generated from a generator, while no data are used in training the generator and in quantization. To this end, we propose data-free adversarial knowledge distillation, which minimizes the maximum distance between the outputs of the teacher and the (quantized) student for any adversarial samples from a generator. To generate adversarial samples similar to the original data, we additionally propose matching statistics from the batch normalization layers for generated data and the original data in the teacher. Furthermore, we show the gain of producing diverse adversarial samples by using multiple generators and multiple students. Our experiments show the state-of-the-art data-free model compression and quantization results for (wide) residual networks and MobileNet on SVHN, CIFAR-10, CIFAR-100, and Tiny-ImageNet datasets. The accuracy losses compared to using the original datasets are shown to be very minimal.
4.NTIRE 2020 Challenge on Real Image Denoising: Dataset, Methods and Results ⬇️
This paper reviews the NTIRE 2020 challenge on real image denoising with focus on the newly introduced dataset, the proposed methods and their results. The challenge is a new version of the previous NTIRE 2019 challenge on real image denoising that was based on the SIDD benchmark. This challenge is based on a newly collected validation and testing image datasets, and hence, named SIDD+. This challenge has two tracks for quantitatively evaluating image denoising performance in (1) the Bayer-pattern rawRGB and (2) the standard RGB (sRGB) color spaces. Each track ~250 registered participants. A total of 22 teams, proposing 24 methods, competed in the final phase of the challenge. The proposed methods by the participating teams represent the current state-of-the-art performance in image denoising targeting real noisy images. The newly collected SIDD+ datasets are publicly available at: this https URL.
5.Sparsely-Labeled Source Assisted Domain Adaptation ⬇️
Domain Adaptation (DA) aims to generalize the classifier learned from the source domain to the target domain. Existing DA methods usually assume that rich labels could be available in the source domain. However, there are usually a large number of unlabeled data but only a few labeled data in the source domain, and how to transfer knowledge from this sparsely-labeled source domain to the target domain is still a challenge, which greatly limits their application in the wild. This paper proposes a novel Sparsely-Labeled Source Assisted Domain Adaptation (SLSA-DA) algorithm to address the challenge with limited labeled source domain samples. Specifically, due to the label scarcity problem, the projected clustering is conducted on both the source and target domains, so that the discriminative structures of data could be leveraged elegantly. Then the label propagation is adopted to propagate the labels from those limited labeled source samples to the whole unlabeled data progressively, so that the cluster labels are revealed correctly. Finally, we jointly align the marginal and conditional distributions to mitigate the cross-domain mismatch problem, and optimize those three procedures iteratively. However, it is nontrivial to incorporate those three procedures into a unified optimization framework seamlessly since some variables to be optimized are implicitly involved in their formulas, thus they could not promote to each other. Remarkably, we prove that the projected clustering and conditional distribution alignment could be reformulated as different expressions, thus the implicit variables are revealed in different optimization steps. As such, the variables related to those three quantities could be optimized in a unified optimization framework and facilitate to each other, to improve the recognition performance obviously.
6.A Sim2Real Deep Learning Approach for the Transformation of Images from Multiple Vehicle-Mounted Cameras to a Semantically Segmented Image in Bird's Eye View ⬇️
Accurate environment perception is essential for automated driving. When using monocular cameras, the distance estimation of elements in the environment poses a major challenge. Distances can be more easily estimated when the camera perspective is transformed to a bird's eye view (BEV). For flat surfaces, Inverse Perspective Mapping (IPM) can accurately transform images to a BEV. Three-dimensional objects such as vehicles and vulnerable road users are distorted by this transformation making it difficult to estimate their position relative to the sensor. This paper describes a methodology to obtain a corrected 360° BEV image given images from multiple vehicle-mounted cameras. The corrected BEV image is segmented into semantic classes and includes a prediction of occluded areas. The neural network approach does not rely on manually labeled data, but is trained on a synthetic dataset in such a way that it generalizes well to real-world data. By using semantically segmented images as input, we reduce the reality gap between simulated and real-world data and are able to show that our method can be successfully applied in the real world. Extensive experiments conducted on the synthetic data demonstrate the superiority of our approach compared to IPM. Source code and datasets are available at this https URL
7.Fast Automatic Visibility Optimization for Thermal Synthetic Aperture Visualization ⬇️
In this article, we describe and validate the first fully automatic parameter optimization for thermal synthetic aperture visualization. It replaces previous manual exploration of the parameter space, which is time consuming and error prone. We prove that the visibility of targets in thermal integral images is proportional to the variance of the targets' image. Since this is invariant to occlusion it represents a suitable objective function for optimization. Our findings have the potential to enable fully autonomous search and recuse operations with camera drones.
8.TSDM: Tracking by SiamRPN++ with a Depth-refiner and a Mask-generator ⬇️
In a generic object tracking, depth (D) information provides informative cues for foreground-background separation and target bounding box regression. However, so far, few trackers have used depth information to play the important role aforementioned due to the lack of a suitable model. In this paper, a RGB-D tracker named TSDM is proposed, which is composed of a Mask-generator (M-g), SiamRPN++ and a Depth-refiner (D-r). The M-g generates the background masks, and updates them as the target 3D position changes. The D-r optimizes the target bounding box estimated by SiamRPN++, based on the spatial depth distribution difference between the target and the surrounding background. Extensive evaluation on the Princeton Tracking Benchmark and the Visual Object Tracking challenge shows that our tracker outperforms the state-of-the-art by a large margin while achieving 23 FPS. In addition, a light-weight variant can run at 31 FPS and thus it is practical for real world applications. Code and models of TSDM are available at this https URL.
9.Introduction of a new Dataset and Method for Detecting and Counting the Pistachios based on Deep Learning ⬇️
Pistachio is a nutritious nut that has many uses in the food industry. Iran is one of its largest producers, and pistachio is considered as a strategic export product for this country. Pistachios are sorted based on the shape of their shell into two categories: Open-mouth and Closed-mouth. The open-mouth pistachios are higher in price, value, and demand than the closed-mouth pistachios. In the countries that are famous in pistachio production and exporting, there are companies that pack the picked pistachios from the trees and make them ready for exporting. As there are differences between the price and the demand of the open-mouth and closed-mouth pistachios, it is considerable for these companies to know precisely how much of these two kinds of pistachios exist in each packed package. In this paper, we have introduced and shared a new dataset of pistachios, which is called Pesteh-Set. Pesteh-Set includes 6 videos with a total length of 164 seconds and 561 moving pistachios. It also contains 423 labeled images that totally include 3927 labeled pistachios. At the first stage, we have used RetinaNet, the deep fully convolutional object detector for detecting the pistachios in the video frames. In the second stage, we introduce our method for counting the open-mouth and closed-mouth pistachios in the videos. Pistachios that move and roll on the transportation line may appear as closed-mouth in some frames and as open-mouth in other frames. With this circumstance, the main challenge of our work is to count these two kinds of pistachios correctly and fast. Our introduced method performs very fast with no need for GPU, and it also achieves good counting results. The computed accuracy of our counting method is 94.75%. Our proposed methods can be remotely performed by using the videos taken from the implemented cameras that could monitor the pistachios.
10.On Vocabulary Reliance in Scene Text Recognition ⬇️
The pursuit of high performance on public benchmarks has been the driving force for research in scene text recognition, and notable progress has been achieved. However, a close investigation reveals a startling fact that the state-of-the-art methods perform well on images with words within vocabulary but generalize poorly to images with words outside vocabulary. We call this phenomenon "vocabulary reliance". In this paper, we establish an analytical framework to conduct an in-depth study on the problem of vocabulary reliance in scene text recognition. Key findings include: (1) Vocabulary reliance is ubiquitous, i.e., all existing algorithms more or less exhibit such characteristic; (2) Attention-based decoders prove weak in generalizing to words outside vocabulary and segmentation-based decoders perform well in utilizing visual features; (3) Context modeling is highly coupled with the prediction layers. These findings provide new insights and can benefit future research in scene text recognition. Furthermore, we propose a simple yet effective mutual learning strategy to allow models of two families (attention-based and segmentation-based) to learn collaboratively. This remedy alleviates the problem of vocabulary reliance and improves the overall scene text recognition performance.
11.RetinaMask: A Face Mask detector ⬇️
Coronavirus disease 2019 has affected the world seriously, because people cannot work as usual in case of infection. One of the effective protection methods for human beings is to wear masks in public areas. Furthermore, many public service providers require customers to use the service only if they wear masks correctly. However, there are only a few research studies about face mask detection. To contribute to public healthcare for human beings, we propose RetinaMask, which is a high-accuracy and efficient face mask detector. The proposed RetinaMask is a one-stage detector, which consists of a feature pyramid network to fuse high-level semantic information with multiple feature maps, and a novel context attention module to focus on detecting face masks. In addition, we also propose a novel cross-class object removal algorithm to reject predictions with low confidences and the high intersection of union. Experimental results show that RetinaMask achieves state-of-the-art results on a public face mask dataset with 2.3% and 1.5% higher than the baseline result in the face and mask detection precision, respectively, and
$11.0%$ and$5.9%$ higher than baseline for recall. Besides, we also explore the possibility of implementing RetinaMask with a light-weighted neural network MobileNet for embedded or mobile devices.
12.Efficient convolutional neural networks with smaller filters for human activity recognition using wearable sensors ⬇️
Recently, human activity recognition (HAR) has been beginning to adopt deep learning to substitute for traditional shallow learning techniques that rely on hand-crafted features. CNNs, in particular, have set latest state-of-the-art on various HAR datasets. However, deep model often requires more computing resources, which limits its applications in embedded HAR. Although many successful methods have been proposed to reduce memory and FLOPs of CNNs, they often involve special network architectures for visual tasks, which are not suitable for deep HAR tasks with time series sensor signals, due to remarkable discrepancy. Therefore, it is necessary to develop lightweight deep models to perform HAR. As filter is the basic unit in constructing CNNs, we must ask whether redesigning smaller filters is applicable for deep HAR. In the paper, inspired by the idea, we proposed a lightweight CNN using re-designed Lego filters for the use of HAR. A set of lower-dimensional filters is used as Lego bricks to be stacked for conventional filters, which does not rely on any special network structure. To our knowledge, this is the first paper that proposes lightweight CNN for HAR in ubiquitous and wearable computing arena. The experiment results on five public HAR datasets, UCI-HAR dataset, OPPORTUNITY dataset, UNIMIB-SHAR dataset, PAMAP2 dataset, and WISDM dataset, indicate that our novel Lego-CNN approach can greatly reduce memory and computation cost over CNN, while maintaining comparable accuracy. We believe that the proposed approach could combine with the existing state-of-the-art HAR architecture and easily deployed onto wearable devices for real HAR applications.
13.Learning Generalized Spoof Cues for Face Anti-spoofing ⬇️
Many existing face anti-spoofing (FAS) methods focus on modeling the decision boundaries for some predefined spoof types. However, the diversity of the spoof samples including the unknown ones hinders the effective decision boundary modeling and leads to weak generalization capability. In this paper, we reformulate FAS in an anomaly detection perspective and propose a residual-learning framework to learn the discriminative live-spoof differences which are defined as the spoof cues. The proposed framework consists of a spoof cue generator and an auxiliary classifier. The generator minimizes the spoof cues of live samples while imposes no explicit constraint on those of spoof samples to generalize well to unseen attacks. In this way, anomaly detection is implicitly used to guide spoof cue generation, leading to discriminative feature learning. The auxiliary classifier serves as a spoof cue amplifier and makes the spoof cues more discriminative. We conduct extensive experiments and the experimental results show the proposed method consistently outperforms the state-of-the-art methods. The code will be publicly available at this https URL.
14.OpenEDS2020: Open Eyes Dataset ⬇️
We present the second edition of OpenEDS dataset, OpenEDS2020, a novel dataset of eye-image sequences captured at a frame rate of 100 Hz under controlled illumination, using a virtual-reality head-mounted display mounted with two synchronized eye-facing cameras. The dataset, which is anonymized to remove any personally identifiable information on participants, consists of 80 participants of varied appearance performing several gaze-elicited tasks, and is divided in two subsets: 1) Gaze Prediction Dataset, with up to 66,560 sequences containing 550,400 eye-images and respective gaze vectors, created to foster research in spatio-temporal gaze estimation and prediction approaches; and 2) Eye Segmentation Dataset, consisting of 200 sequences sampled at 5 Hz, with up to 29,500 images, of which 5% contain a semantic segmentation label, devised to encourage the use of temporal information to propagate labels to contiguous frames. Baseline experiments have been evaluated on OpenEDS2020, one for each task, with average angular error of 5.37 degrees when performing gaze prediction on 1 to 5 frames into the future, and a mean intersection over union score of 84.1% for semantic segmentation. As its predecessor, OpenEDS dataset, we anticipate that this new dataset will continue creating opportunities to researchers in eye tracking, machine learning and computer vision communities, to advance the state of the art for virtual reality applications. The dataset is available for download upon request at this http URL.
15.Point Cloud Completion by Skip-attention Network with Hierarchical Folding ⬇️
Point cloud completion aims to infer the complete geometries for missing regions of 3D objects from incomplete ones. Previous methods usually predict the complete point cloud based on the global shape representation extracted from the incomplete input. However, the global representation often suffers from the information loss of structure details on local regions of incomplete point cloud. To address this problem, we propose Skip-Attention Network (SA-Net) for 3D point cloud completion. Our main contributions lie in the following two-folds. First, we propose a skip-attention mechanism to effectively exploit the local structure details of incomplete point clouds during the inference of missing parts. The skip-attention mechanism selectively conveys geometric information from the local regions of incomplete point clouds for the generation of complete ones at different resolutions, where the skip-attention reveals the completion process in an interpretable way. Second, in order to fully utilize the selected geometric information encoded by skip-attention mechanism at different resolutions, we propose a novel structure-preserving decoder with hierarchical folding for complete shape generation. The hierarchical folding preserves the structure of complete point cloud generated in upper layer by progressively detailing the local regions, using the skip-attentioned geometry at the same resolution. We conduct comprehensive experiments on ShapeNet and KITTI datasets, which demonstrate that the proposed SA-Net outperforms the state-of-the-art point cloud completion methods.
16.Where am I looking at? Joint Location and Orientation Estimation by Cross-View Matching ⬇️
Cross-view geo-localization is the problem of estimating the position and orientation (latitude, longitude and azimuth angle) of a camera at ground level given a large-scale database of geo-tagged aerial (e.g., satellite) images. Existing approaches treat the task as a pure location estimation problem by learning discriminative feature descriptors, but neglect orientation alignment. It is well-recognized that knowing the orientation between ground and aerial images can significantly reduce matching ambiguity between these two views, especially when the ground-level images have a limited Field of View (FoV) instead of a full field-of-view panorama. Therefore, we design a Dynamic Similarity Matching network to estimate cross-view orientation alignment during localization. In particular, we address the cross-view domain gap by applying a polar transform to the aerial images to approximately align the images up to an unknown azimuth angle. Then, a two-stream convolutional network is used to learn deep features from the ground and polar-transformed aerial images. Finally, we obtain the orientation by computing the correlation between cross-view features, which also provides a more accurate measure of feature similarity, improving location recall. Experiments on standard datasets demonstrate that our method significantly improves state-of-the-art performance. Remarkably, we improve the top-1 location recall rate on the CVUSA dataset by a factor of 1.5x for panoramas with known orientation, by a factor of 3.3x for panoramas with unknown orientation, and by a factor of 6x for 180-degree FoV images with unknown orientation.
17.Synchronous Bidirectional Learning for Multilingual Lip Reading ⬇️
Lip reading has received increasing attention in recent years. This paper focuses on the synergy of multilingual lip reading. There are more than 7,000 languages in the world, which implies that it is impractical to train separate lip reading models by collecting large-scale data per language. Although each language has its own linguistic and pronunciation features, the lip movements of all languages share similar patterns. Based on this idea, in this paper, we try to explore the synergized learning of multilingual lip reading, and further propose a synchronous bidirectional learning(SBL) framework for effective synergy of multilingual lip reading. Firstly, we introduce the phonemes as our modeling units for the multilingual setting. Similar phoneme always leads to similar visual patterns. The multilingual setting would increase both the quantity and the diversity of each phoneme shared among different languages. So the learning for the multilingual target should bring improvement to the prediction of phonemes. Then, a SBL block is proposed to infer the target unit when given its previous and later context. The rules for each specific language which the model itself judges to be is learned in this fill-in-the-blank manner. To make the learning process more targeted at each particular language, we introduce an extra task of predicting the language identity in the learning process. Finally, we perform a thorough comparison on LRW (English) and LRW-1000(Mandarin). The results outperform the existing state of the art by a large margin, and show the promising benefits from the synergized learning of different languages.
18.SurfelGAN: Synthesizing Realistic Sensor Data for Autonomous Driving ⬇️
Autonomous driving system development is critically dependent on the ability to replay complex and diverse traffic scenarios in simulation. In such scenarios, the ability to accurately simulate the vehicle sensors such as cameras, lidar or radar is essential. However, current sensor simulators leverage gaming engines such as Unreal or Unity, requiring manual creation of environments, objects and material properties. Such approaches have limited scalability and fail to produce realistic approximations of camera, lidar, and radar data without significant additional work.
In this paper, we present a simple yet effective approach to generate realistic scenario sensor data, based only on a limited amount of lidar and camera data collected by an autonomous vehicle. Our approach uses texture-mapped surfels to efficiently reconstruct the scene from an initial vehicle pass or set of passes, preserving rich information about object 3D geometry and appearance, as well as the scene conditions. We then leverage a SurfelGAN network to reconstruct realistic camera images for novel positions and orientations of the self-driving vehicle and moving objects in the scene. We demonstrate our approach on the Waymo Open Dataset and show that it can synthesize realistic camera data for simulated scenarios. We also create a novel dataset that contains cases in which two self-driving vehicles observe the same scene at the same time. We use this dataset to provide additional evaluation and demonstrate the usefulness of our SurfelGAN model.
19.Projection & Probability-Driven Black-Box Attack ⬇️
Generating adversarial examples in a black-box setting retains a significant challenge with vast practical application prospects. In particular, existing black-box attacks suffer from the need for excessive queries, as it is non-trivial to find an appropriate direction to optimize in the high-dimensional space. In this paper, we propose Projection & Probability-driven Black-box Attack (PPBA) to tackle this problem by reducing the solution space and providing better optimization. For reducing the solution space, we first model the adversarial perturbation optimization problem as a process of recovering frequency-sparse perturbations with compressed sensing, under the setting that random noise in the low-frequency space is more likely to be adversarial. We then propose a simple method to construct a low-frequency constrained sensing matrix, which works as a plug-and-play projection matrix to reduce the dimensionality. Such a sensing matrix is shown to be flexible enough to be integrated into existing methods like NES and Bandits$_{TD}$. For better optimization, we perform a random walk with a probability-driven strategy, which utilizes all queries over the whole progress to make full use of the sensing matrix for a less query budget. Extensive experiments show that our method requires at most 24% fewer queries with a higher attack success rate compared with state-of-the-art approaches. Finally, the attack method is evaluated on the real-world online service, i.e., Google Cloud Vision API, which further demonstrates our practical potentials.
20.One-Shot Object Detection without Fine-Tuning ⬇️
Deep learning has revolutionized object detection thanks to large-scale datasets, but their object categories are still arguably very limited. In this paper, we attempt to enrich such categories by addressing the one-shot object detection problem, where the number of annotated training examples for learning an unseen class is limited to one. We introduce a two-stage model consisting of a first stage Matching-FCOS network and a second stage Structure-Aware Relation Module, the combination of which integrates metric learning with an anchor-free Faster R-CNN-style detection pipeline, eventually eliminating the need to fine-tune on the support images. We also propose novel training strategies that effectively improve detection performance. Extensive quantitative and qualitative evaluations were performed and our method exceeds the state-of-the-art one-shot performance consistently on multiple datasets.
21.Text Synopsis Generation for Egocentric Videos ⬇️
Mass utilization of body-worn cameras has led to a huge corpus of available egocentric video. Existing video summarization algorithms can accelerate browsing such videos by selecting (visually) interesting shots from them. Nonetheless, since the system user still has to watch the summary videos, browsing large video databases remain a challenge. Hence, in this work, we propose to generate a textual synopsis, consisting of a few sentences describing the most important events in a long egocentric videos. Users can read the short text to gain insight about the video, and more importantly, efficiently search through the content of a large video database using text queries. Since egocentric videos are long and contain many activities and events, using video-to-text algorithms results in thousands of descriptions, many of which are incorrect. Therefore, we propose a multi-task learning scheme to simultaneously generate descriptions for video segments and summarize the resulting descriptions in an end-to-end fashion. We Input a set of video shots and the network generates a text description for each shot. Next, visual-language content matching unit that is trained with a weakly supervised objective, identifies the correct descriptions. Finally, the last component of our network, called purport network, evaluates the descriptions all together to select the ones containing crucial information. Out of thousands of descriptions generated for the video, a few informative sentences are returned to the user. We validate our framework on the challenging UT Egocentric video dataset, where each video is between 3 to 5 hours long, associated with over 3000 textual descriptions on average. The generated textual summaries, including only 5 percent (or less) of the generated descriptions, are compared to groundtruth summaries in text domain using well-established metrics in natural language processing.
22.A Gaussian Process Upsampling Model for Improvements in Optical Character Recognition ⬇️
Optical Character Recognition and extraction is a key tool in the automatic evaluation of documents in a financial context. However, the image data provided to automated systems can have unreliable quality, and can be inherently low-resolution or downsampled and compressed by a transmitting program. In this paper, we illustrate the efficacy of a Gaussian Process upsampling model for the purposes of improving OCR and extraction through upsampling low resolution documents.
23.Recognizing Magnification Levels in Microscopic Snapshots ⬇️
Recent advances in digital imaging has transformed computer vision and machine learning to new tools for analyzing pathology images. This trend could automate some of the tasks in the diagnostic pathology and elevate the pathologist workload. The final step of any cancer diagnosis procedure is performed by the expert pathologist. These experts use microscopes with high level of optical magnification to observe minute characteristics of the tissue acquired through biopsy and fixed on glass slides. Switching between different magnifications, and finding the magnification level at which they identify the presence or absence of malignant tissues is important. As the majority of pathologists still use light microscopy, compared to digital scanners, in many instance a mounted camera on the microscope is used to capture snapshots from significant field-of-views. Repositories of such snapshots usually do not contain the magnification information. In this paper, we extract deep features of the images available on TCGA dataset with known magnification to train a classifier for magnification recognition. We compared the results with LBP, a well-known handcrafted feature extraction method. The proposed approach achieved a mean accuracy of 96% when a multi-layer perceptron was trained as a classifier.
24.Effective Data Fusion with Generalized Vegetation Index: Evidence from Land Cover Segmentation in Agriculture ⬇️
How can we effectively leverage the domain knowledge from remote sensing to better segment agriculture land cover from satellite images? In this paper, we propose a novel, model-agnostic, data-fusion approach for vegetation-related computer vision tasks. Motivated by the various Vegetation Indices (VIs), which are introduced by domain experts, we systematically reviewed the VIs that are widely used in remote sensing and their feasibility to be incorporated in deep neural networks. To fully leverage the Near-Infrared channel, the traditional Red-Green-Blue channels, and Vegetation Index or its variants, we propose a Generalized Vegetation Index (GVI), a lightweight module that can be easily plugged into many neural network architectures to serve as an additional information input. To smoothly train models with our GVI, we developed an Additive Group Normalization (AGN) module that does not require extra parameters of the prescribed neural networks. Our approach has improved the IoUs of vegetation-related classes by 0.9-1.3 percent and consistently improves the overall mIoU by 2 percent on our baseline.
25.Neural Object Learning for 6D Pose Estimation Using a Few Cluttered Images ⬇️
Recent methods for 6D pose estimation of objects assume either textured 3D models or real images that cover the entire range of target poses. However, it is difficult to obtain textured 3D models and annotate the poses of objects in real scenarios. This paper proposes a method, Neural Object Learning (NOL), that creates synthetic images of objects in arbitrary poses by combining only a few observations from cluttered images. A novel refinement step is proposed to align inaccurate poses of objects in source images, which results in better quality images. Evaluations performed on two public datasets show that the rendered images created by NOL lead to state-of-the-art performance in comparison to methods that use 10 times the number of real images. Evaluations on our new dataset show multiple objects can be trained and recognized simultaneously using a sequence of a fixed scene.
26.Regularized Pooling ⬇️
In convolutional neural networks (CNNs), pooling operations play important roles such as dimensionality reduction and deformation compensation. In general, max pooling, which is the most widely used operation for local pooling, is performed independently for each kernel. However, the deformation may be spatially smooth over the neighboring kernels. This means that max pooling is too flexible to compensate for actual deformations. In other words, its excessive flexibility risks canceling the essential spatial differences between classes. In this paper, we propose regularized pooling, which enables the value selection direction in the pooling operation to be spatially smooth across adjacent kernels so as to compensate only for actual deformations. The results of experiments on handwritten character images and texture images showed that regularized pooling not only improves recognition accuracy but also accelerates the convergence of learning compared with conventional pooling operations.
27.Source-Relaxed Domain Adaptation for Image Segmentation ⬇️
Domain adaptation (DA) has drawn high interests for its capacity to adapt a model trained on labeled source data to perform well on unlabeled or weakly labeled target data from a different domain. Most common DA techniques require the concurrent access to the input images of both the source and target domains. However, in practice, it is common that the source images are not available in the adaptation phase. This is a very frequent DA scenario in medical imaging, for instance, when the source and target images come from different clinical sites. We propose a novel formulation for adapting segmentation networks, which relaxes such a constraint. Our formulation is based on minimizing a label-free entropy loss defined over target-domain data, which we further guide with a domain invariant prior on the segmentation regions. Many priors can be used, derived from anatomical information. Here, a class-ratio prior is learned via an auxiliary network and integrated in the form of a Kullback-Leibler (KL) divergence in our overall loss function. We show the effectiveness of our prior-aware entropy minimization in adapting spine segmentation across different MRI modalities. Our method yields comparable results to several state-of-the-art adaptation techniques, even though is has access to less information, the source images being absent in the adaptation phase. Our straight-forward adaptation strategy only uses one network, contrary to popular adversarial techniques, which cannot perform without the presence of the source images. Our framework can be readily used with various priors and segmentation problems.
28.Preprint: Using RF-DNA Fingerprints To Classify OFDM Transmitters Under Rayleigh Fading Conditions ⬇️
The Internet of Things (IoT) is a collection of Internet connected devices capable of interacting with the physical world and computer systems. It is estimated that the IoT will consist of approximately fifty billion devices by the year 2020. In addition to the sheer numbers, the need for IoT security is exacerbated by the fact that many of the edge devices employ weak to no encryption of the communication link. It has been estimated that almost 70% of IoT devices use no form of encryption. Previous research has suggested the use of Specific Emitter Identification (SEI), a physical layer technique, as a means of augmenting bit-level security mechanism such as encryption. The work presented here integrates a Nelder-Mead based approach for estimating the Rayleigh fading channel coefficients prior to the SEI approach known as RF-DNA fingerprinting. The performance of this estimator is assessed for degrading signal-to-noise ratio and compared with least square and minimum mean squared error channel estimators. Additionally, this work presents classification results using RF-DNA fingerprints that were extracted from received signals that have undergone Rayleigh fading channel correction using Minimum Mean Squared Error (MMSE) equalization. This work also performs radio discrimination using RF-DNA fingerprints generated from the normalized magnitude-squared and phase response of Gabor coefficients as well as two classifiers. Discrimination of four 802.11a Wi-Fi radios achieves an average percent correct classification of 90% or better for signal-to-noise ratios of 18 and 21 dB or greater using a Rayleigh fading channel comprised of two and five paths, respectively.
29.A Hybrid Method for Training Convolutional Neural Networks ⬇️
Artificial Intelligence algorithms have been steadily increasing in popularity and usage. Deep Learning, allows neural networks to be trained using huge datasets and also removes the need for human extracted features, as it automates the feature learning process. In the hearth of training deep neural networks, such as Convolutional Neural Networks, we find backpropagation, that by computing the gradient of the loss function with respect to the weights of the network for a given input, it allows the weights of the network to be adjusted to better perform in the given task. In this paper, we propose a hybrid method that uses both backpropagation and evolutionary strategies to train Convolutional Neural Networks, where the evolutionary strategies are used to help to avoid local minimas and fine-tune the weights, so that the network achieves higher accuracy results. We show that the proposed hybrid method is capable of improving upon regular training in the task of image classification in CIFAR-10, where a VGG16 model was used and the final test results increased 0.61%, in average, when compared to using only backpropagation.
30.Multi-Phase Cross-modal Learning for Noninvasive Gene Mutation Prediction in Hepatocellular Carcinoma ⬇️
Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer and the fourth most common cause of cancer-related death worldwide. Understanding the underlying gene mutations in HCC provides great prognostic value for treatment planning and targeted therapy. Radiogenomics has revealed an association between non-invasive imaging features and molecular genomics. However, imaging feature identification is laborious and error-prone. In this paper, we propose an end-to-end deep learning framework for mutation prediction in APOB, COL11A1 and ATRX genes using multiphasic CT scans. Considering intra-tumour heterogeneity (ITH) in HCC, multi-region sampling technology is implemented to generate the dataset for experiments. Experimental results demonstrate the effectiveness of the proposed model.
31.Hypergraph Learning for Identification of COVID-19 with CT Imaging ⬇️
The coronavirus disease, named COVID-19, has become the largest global public health crisis since it started in early 2020. CT imaging has been used as a complementary tool to assist early screening, especially for the rapid identification of COVID-19 cases from community acquired pneumonia (CAP) cases. The main challenge in early screening is how to model the confusing cases in the COVID-19 and CAP groups, with very similar clinical manifestations and imaging features. To tackle this challenge, we propose an Uncertainty Vertex-weighted Hypergraph Learning (UVHL) method to identify COVID-19 from CAP using CT images. In particular, multiple types of features (including regional features and radiomics features) are first extracted from CT image for each case. Then, the relationship among different cases is formulated by a hypergraph structure, with each case represented as a vertex in the hypergraph. The uncertainty of each vertex is further computed with an uncertainty score measurement and used as a weight in the hypergraph. Finally, a learning process of the vertex-weighted hypergraph is used to predict whether a new testing case belongs to COVID-19 or not. Experiments on a large multi-center pneumonia dataset, consisting of 2,148 COVID-19 cases and 1,182 CAP cases from five hospitals, are conducted to evaluate the performance of the proposed method. Results demonstrate the effectiveness and robustness of our proposed method on the identification of COVID-19 in comparison to state-of-the-art methods.
32.Convolutional Sparse Support Estimator Based Covid-19 Recognition from X-ray Images ⬇️
Coronavirus disease (Covid-19) has been the main agenda of the whole world since it came in sight in December 2019. It has already caused thousands of causalities and infected several millions worldwide. Any technological tool that can be provided to healthcare practitioners to save time, effort, and possibly lives has crucial importance. The main tools practitioners currently use to diagnose Covid-19 are Reverse Transcription-Polymerase Chain reaction (RT-PCR) and Computed Tomography (CT), which require significant time, resources and acknowledged experts. X-ray imaging is a common and easily accessible tool that has great potential for Covid-19 diagnosis. In this study, we propose a novel approach for Covid-19 recognition from chest X-ray images. Despite the importance of the problem, recent studies in this domain produced not so satisfactory results due to the limited datasets available for training. Recall that Deep Learning techniques can generally provide state-of-the-art performance in many classification tasks when trained properly over large datasets, such data scarcity can be a crucial obstacle when using them for Covid-19 detection. Alternative approaches such as representation-based classification (collaborative or sparse representation) might provide satisfactory performance with limited size datasets, but they generally fall short in performance or speed compared to Machine Learning methods. To address this deficiency, Convolution Support Estimation Network (CSEN) has recently been proposed as a bridge between model-based and Deep Learning approaches by providing a non-iterative real-time mapping from query sample to ideally sparse representation coefficient' support, which is critical information for class decision in representation based techniques.
33.DeepHist: Differentiable Joint and Color Histogram Layers for Image-to-Image Translation ⬇️
We present the DeepHist - a novel Deep Learning framework for augmenting a network by histogram layers and demonstrate its strength by addressing image-to-image translation problems. Specifically, given an input image and a reference color distribution we aim to generate an output image with the structural appearance (content) of the input (source) yet with the colors of the reference. The key idea is a new technique for a differentiable construction of joint and color histograms of the output images. We further define a color distribution loss based on the Earth Mover's Distance between the output's and the reference's color histograms and a Mutual Information loss based on the joint histograms of the source and the output images. Promising results are shown for the tasks of color transfer, image colorization and edges
$\rightarrow$ photo, where the color distribution of the output image is controlled. Comparison to Pix2Pix and CyclyGANs are shown.
34.Compressive sensing with un-trained neural networks: Gradient descent finds the smoothest approximation ⬇️
Un-trained convolutional neural networks have emerged as highly successful tools for image recovery and restoration. They are capable of solving standard inverse problems such as denoising and compressive sensing with excellent results by simply fitting a neural network model to measurements from a single image or signal without the need for any additional training data. For some applications, this critically requires additional regularization in the form of early stopping the optimization. For signal recovery from a few measurements, however, un-trained convolutional networks have an intriguing self-regularizing property: Even though the network can perfectly fit any image, the network recovers a natural image from few measurements when trained with gradient descent until convergence. In this paper, we provide numerical evidence for this property and study it theoretically. We show that---without any further regularization---an un-trained convolutional neural network can approximately reconstruct signals and images that are sufficiently structured, from a near minimal number of random measurements.
35.Beyond CNNs: Exploiting Further Inherent Symmetries in Medical Images for Segmentation ⬇️
Automatic tumor segmentation is a crucial step in medical image analysis for computer-aided diagnosis. Although the existing methods based on convolutional neural networks (CNNs) have achieved the state-of-the-art performance, many challenges still remain in medical tumor segmentation. This is because regular CNNs can only exploit translation invariance, ignoring further inherent symmetries existing in medical images such as rotations and reflections. To mitigate this shortcoming, we propose a novel group equivariant segmentation framework by encoding those inherent symmetries for learning more precise representations. First, kernel-based equivariant operations are devised on every orientation, which can effectively address the gaps of learning symmetries in existing approaches. Then, to keep segmentation networks globally equivariant, we design distinctive group layers with layerwise symmetry constraints. By exploiting further symmetries, novel segmentation CNNs can dramatically reduce the sample complexity and the redundancy of filters (by roughly 2/3) over regular CNNs. More importantly, based on our novel framework, we show that a newly built GER-UNet outperforms its regular CNN-based counterpart and the state-of-the-art segmentation methods on real-world clinical data. Specifically, the group layers of our segmentation framework can be seamlessly integrated into any popular CNN-based segmentation architectures.
36.Is an Affine Constraint Needed for Affine Subspace Clustering? ⬇️
Subspace clustering methods based on expressing each data point as a linear combination of other data points have achieved great success in computer vision applications such as motion segmentation, face and digit clustering. In face clustering, the subspaces are linear and subspace clustering methods can be applied directly. In motion segmentation, the subspaces are affine and an additional affine constraint on the coefficients is often enforced. However, since affine subspaces can always be embedded into linear subspaces of one extra dimension, it is unclear if the affine constraint is really necessary. This paper shows, both theoretically and empirically, that when the dimension of the ambient space is high relative to the sum of the dimensions of the affine subspaces, the affine constraint has a negligible effect on clustering performance. Specifically, our analysis provides conditions that guarantee the correctness of affine subspace clustering methods both with and without the affine constraint, and shows that these conditions are satisfied for high-dimensional data. Underlying our analysis is the notion of affinely independent subspaces, which not only provides geometrically interpretable correctness conditions, but also clarifies the relationships between existing results for affine subspace clustering.
37.Synergistic Learning of Lung Lobe Segmentation and Hierarchical Multi-Instance Classification for Automated Severity Assessment of COVID-19 in CT Images ⬇️
Understanding chest CT imaging of the coronavirus disease 2019 (COVID-19) will help detect infections early and assess the disease progression. Especially, automated severity assessment of COVID-19 in CT images plays an essential role in identifying cases that are in great need of intensive clinical care. However, it is often challenging to accurately assess the severity of this disease in CT images, due to small infection regions in the lungs, similar imaging biomarkers, and large inter-case variations. To this end, we propose a synergistic learning framework for automated severity assessment of COVID-19 in 3D CT images, by jointly performing lung lobe segmentation and multi-instance classification. Considering that only a few infection regions in a CT image are related to the severity assessment, we first represent each input image by a bag that contains a set of 2D image patches (with each one cropped from a specific slice). A multi-task multi-instance deep network (called M2UNet) is then developed to assess the severity of COVID-19 patients and segment the lung lobe simultaneously. Our M2UNet consists of a patch-level encoder, a segmentation sub-network for lung lobe segmentation, and a classification sub-network for severity assessment (with a unique hierarchical multi-instance learning strategy). Here, the context information provided by segmentation can be implicitly employed to improve the performance of severity assessment. Extensive experiments were performed on a real COVID-19 CT image dataset consisting of 666 chest CT images, with results suggesting the effectiveness of our proposed method compared to several state-of-the-art methods.
38.Y-Net for Chest X-Ray Preprocessing: Simultaneous Classification of Geometry and Segmentation of Annotations ⬇️
Over the last decade, convolutional neural networks (CNNs) have emerged as the leading algorithms in image classification and segmentation. Recent publication of large medical imaging databases have accelerated their use in the biomedical arena. While training data for photograph classification benefits from aggressive geometric augmentation, medical diagnosis -- especially in chest radiographs -- depends more strongly on feature location. Diagnosis classification results may be artificially enhanced by reliance on radiographic annotations. This work introduces a general pre-processing step for chest x-ray input into machine learning algorithms. A modified Y-Net architecture based on the VGG11 encoder is used to simultaneously learn geometric orientation (similarity transform parameters) of the chest and segmentation of radiographic annotations. Chest x-rays were obtained from published databases. The algorithm was trained with 1000 manually labeled images with augmentation. Results were evaluated by expert clinicians, with acceptable geometry in 95.8% and annotation mask in 96.2% (n=500), compared to 27.0% and 34.9% respectively in control images (n=241). We hypothesize that this pre-processing step will improve robustness in future diagnostic algorithms.
39.Blind Backdoors in Deep Learning Models ⬇️
We investigate a new method for injecting backdoors into machine learning models, based on poisoning the loss computation in the model-training code. Our attack is blind: the attacker cannot modify the training data, nor observe the execution of his code, nor access the resulting model. We develop a new technique for blind backdoor training using multi-objective optimization to achieve high accuracy on both the main and backdoor tasks while evading all known defenses. We then demonstrate the efficacy of the blind attack with new classes of backdoors strictly more powerful than those in prior literature: single-pixel backdoors in ImageNet models, backdoors that switch the model to a different, complex task, and backdoors that do not require inference-time input modifications. Finally, we discuss defenses.
40.MLGaze: Machine Learning-Based Analysis of Gaze Error Patterns in Consumer Eye Tracking Systems ⬇️
Analyzing the gaze accuracy characteristics of an eye tracker is a critical task as its gaze data is frequently affected by non-ideal operating conditions in various consumer eye tracking applications. In this study, gaze error patterns produced by a commercial eye tracking device were studied with the help of machine learning algorithms, such as classifiers and regression models. Gaze data were collected from a group of participants under multiple conditions that commonly affect eye trackers operating on desktop and handheld platforms. These conditions (referred here as error sources) include user distance, head pose, and eye-tracker pose variations, and the collected gaze data were used to train the classifier and regression models. It was seen that while the impact of the different error sources on gaze data characteristics were nearly impossible to distinguish by visual inspection or from data statistics, machine learning models were successful in identifying the impact of the different error sources and predicting the variability in gaze error levels due to these conditions. The objective of this study was to investigate the efficacy of machine learning methods towards the detection and prediction of gaze error patterns, which would enable an in-depth understanding of the data quality and reliability of eye trackers under unconstrained operating conditions. Coding resources for all the machine learning methods adopted in this study were included in an open repository named MLGaze to allow researchers to replicate the principles presented here using data from their own eye trackers.
41.Federated Generative Adversarial Learning ⬇️
This work studies training generative adversarial networks under the federated learning setting. Generative adversarial networks (GANs) have achieved advancement in various real-world applications, such as image editing, style transfer, scene generations, etc. However, like other deep learning models, GANs are also suffering from data limitation problems in real cases. To boost the performance of GANs in target tasks, collecting images as many as possible from different sources becomes not only important but also essential. For example, to build a robust and accurate bio-metric verification system, huge amounts of images might be collected from surveillance cameras, and/or uploaded from cellphones by users accepting agreements. In an ideal case, utilize all those data uploaded from public and private devices for model training is straightforward. Unfortunately, in the real scenarios, this is hard due to a few reasons. At first, some data face the serious concern of leakage, and therefore it is prohibitive to upload them to a third-party server for model training; at second, the images collected by different kinds of devices, probably have distinctive biases due to various factors,
$\textit{e.g.}$ , collector preferences, geo-location differences, which is also known as "domain shift". To handle those problems, we propose a novel generative learning scheme utilizing a federated learning framework. Following the configuration of federated learning, we conduct model training and aggregation on one center and a group of clients. Specifically, our method learns the distributed generative models in clients, while the models trained in each client are fused into one unified and versatile model in the center. We perform extensive experiments to compare different federation strategies, and empirically examine the effectiveness of federation under different levels of parallelism and data skewness.
42.ProSelfLC: Progressive Self Label Correction for Target Revising in Label Noise ⬇️
In this work, we address robust deep learning under label noise (semi-supervised learning) from the perspective of target revising. We make three main contributions. First, we present a comprehensive mathematical study on existing target modification techniques, including Pseudo-Label [1], label smoothing [2], bootstrapping [3], knowledge distillation [4], confidence penalty [5], and joint optimisation [6]. Consequently, we reveal their relationships and drawbacks. Second, we propose ProSelfLC, a progressive and adaptive self label correction method, endorsed by learning time and predictive confidence. It addresses the disadvantages of existing algorithms and embraces many practical merits: (1) It is end-to-end trainable; (2) Given an example, ProSelfLC has the ability to revise an one-hot target by adding the information about its similarity structure, and correcting its semantic class; (3) No auxiliary annotations, or extra learners are required. Our proposal is designed according to the well-known expertise: deep neural networks learn simple meaningful patterns before fitting noisy patterns [7-9], and entropy regularisation principle [10, 11]. Third, label smoothing, confidence penalty and naive label correction perform on par with the state-of-the-art in our implementation. This probably indicates they were not benchmarked properly in prior work. Furthermore, our ProSelfLC outperforms them significantly.
43.Planning from Images with Deep Latent Gaussian Process Dynamics ⬇️
Planning is a powerful approach to control problems with known environment dynamics. In unknown environments the agent needs to learn a model of the system dynamics to make planning applicable. This is particularly challenging when the underlying states are only indirectly observable through images. We propose to learn a deep latent Gaussian process dynamics (DLGPD) model that learns low-dimensional system dynamics from environment interactions with visual observations. The method infers latent state representations from observations using neural networks and models the system dynamics in the learned latent space with Gaussian processes. All parts of the model can be trained jointly by optimizing a lower bound on the likelihood of transitions in image space. We evaluate the proposed approach on the pendulum swing-up task while using the learned dynamics model for planning in latent space in order to solve the control problem. We also demonstrate that our method can quickly adapt a trained agent to changes in the system dynamics from just a few rollouts. We compare our approach to a state-of-the-art purely deep learning based method and demonstrate the advantages of combining Gaussian processes with deep learning for data efficiency and transfer learning.
44.A Hand Motion-guided Articulation and Segmentation Estimation ⬇️
In this paper, we present a method for simultaneous articulation model estimation and segmentation of an articulated object in RGB-D images using human hand motion. Our method uses the hand motion in the processes of the initial articulation model estimation, ICP-based model parameter optimization, and region selection of the target object. The hand motion gives an initial guess of the articulation model: prismatic or revolute joint. The method estimates the joint parameters by aligning the RGB-D images with the constraint of the hand motion. Finally, the target regions are selected from the cluster regions which move symmetrically along with the articulation model. Our experimental results show the robustness of the proposed method for the various objects.
45.Learning to Segment Actions from Observation and Narration ⬇️
We apply a generative segmental model of task structure, guided by narration, to action segmentation in video. We focus on unsupervised and weakly-supervised settings where no action labels are known during training. Despite its simplicity, our model performs competitively with previous work on a dataset of naturalistic instructional videos. Our model allows us to vary the sources of supervision used in training, and we find that both task structure and narrative language provide large benefits in segmentation quality.