1.Focus on defocus: bridging the synthetic to real domain gap for depth estimation ⬇️
Data-driven depth estimation methods struggle with the generalization outside their training scenes due to the immense variability of the real-world scenes. This problem can be partially addressed by utilising synthetically generated images, but closing the synthetic-real domain gap is far from trivial. In this paper, we tackle this issue by using domain invariant defocus blur as direct supervision. We leverage defocus cues by using a permutation invariant convolutional neural network that encourages the network to learn from the differences between images with a different point of focus. Our proposed network uses the defocus map as an intermediate supervisory signal. We are able to train our model completely on synthetic data and directly apply it to a wide range of real-world images. We evaluate our model on synthetic and real datasets, showing compelling generalization results and state-of-the-art depth prediction.
2.CIAGAN: Conditional Identity Anonymization Generative Adversarial Networks ⬇️
The unprecedented increase in the usage of computer vision technology in society goes hand in hand with an increased concern in data privacy. In many real-world scenarios like people tracking or action recognition, it is important to be able to process the data while taking careful consideration in protecting people's identity. We propose and develop CIAGAN, a model for image and video anonymization based on conditional generative adversarial networks. Our model is able to remove the identifying characteristics of faces and bodies while producing high-quality images and videos that can be used for any computer vision task, such as detection or tracking. Unlike previous methods, we have full control over the de-identification (anonymization) procedure, ensuring both anonymization as well as diversity. We compare our method to several baselines and achieve state-of-the-art results.
3.Ultrasound Video Summarization using Deep Reinforcement Learning ⬇️
Video is an essential imaging modality for diagnostics, e.g. in ultrasound imaging, for endoscopy, or movement assessment. However, video hasn't received a lot of attention in the medical image analysis community. In the clinical practice, it is challenging to utilise raw diagnostic video data efficiently as video data takes a long time to process, annotate or audit. In this paper we introduce a novel, fully automatic video summarization method that is tailored to the needs of medical video data. Our approach is framed as reinforcement learning problem and produces agents focusing on the preservation of important diagnostic information. We evaluate our method on videos from fetal ultrasound screening, where commonly only a small amount of the recorded data is used diagnostically. We show that our method is superior to alternative video summarization methods and that it preserves essential information required by clinical diagnostic standards.
4.Differentiable Mapping Networks: Learning Structured Map Representations for Sparse Visual Localization ⬇️
Mapping and localization, preferably from a small number of observations, are fundamental tasks in robotics. We address these tasks by combining spatial structure (differentiable mapping) and end-to-end learning in a novel neural network architecture: the Differentiable Mapping Network (DMN). The DMN constructs a spatially structured view-embedding map and uses it for subsequent visual localization with a particle filter. Since the DMN architecture is end-to-end differentiable, we can jointly learn the map representation and localization using gradient descent. We apply the DMN to sparse visual localization, where a robot needs to localize in a new environment with respect to a small number of images from known viewpoints. We evaluate the DMN using simulated environments and a challenging real-world Street View dataset. We find that the DMN learns effective map representations for visual localization. The benefit of spatial structure increases with larger environments, more viewpoints for mapping, and when training data is scarce. Project website: this http URL
5.Toward Automated Classroom Observation: Multimodal Machine Learning to Estimate CLASS Positive Climate and Negative Climate ⬇️
In this work we present a multi-modal machine learning-based system, which we call ACORN, to analyze videos of school classrooms for the Positive Climate (PC) and Negative Climate (NC) dimensions of the CLASS observation protocol that is widely used in educational research. ACORN uses convolutional neural networks to analyze spectral audio features, the faces of teachers and students, and the pixels of each image frame, and then integrates this information over time using Temporal Convolutional Networks. The audiovisual ACORN's PC and NC predictions have Pearson correlations of
$0.55$ and$0.63$ with ground-truth scores provided by expert CLASS coders on the UVA Toddler dataset (cross-validation on$n=300$ 15-min video segments), and a purely auditory ACORN predicts PC and NC with correlations of$0.36$ and$0.41$ on the MET dataset (test set of$n=2000$ videos segments). These numbers are similar to inter-coder reliability of human coders. Finally, using Graph Convolutional Networks we make early strides (AUC=$0.70$) toward predicting the specific moments (45-90sec clips) when the PC is particularly weak/strong. Our findings inform the design of automatic classroom observation and also more general video activity recognition and summary recognition systems.
6.RoadText-1K: Text Detection & Recognition Dataset for Driving Videos ⬇️
Perceiving text is crucial to understand semantics of outdoor scenes and hence is a critical requirement to build intelligent systems for driver assistance and self-driving. Most of the existing datasets for text detection and recognition comprise still images and are mostly compiled keeping text in mind. This paper introduces a new "RoadText-1K" dataset for text in driving videos. The dataset is 20 times larger than the existing largest dataset for text in videos. Our dataset comprises 1000 video clips of driving without any bias towards text and with annotations for text bounding boxes and transcriptions in every frame. State of the art methods for text detection, recognition and tracking are evaluated on the new dataset and the results signify the challenges in unconstrained driving videos compared to existing datasets. This suggests that RoadText-1K is suited for research and development of reading systems, robust enough to be incorporated into more complex downstream tasks like driver assistance and self-driving. The dataset can be found at this http URL
7.Built Infrastructure Monitoring and Inspection Using UAVs and Vision-based Algorithms ⬇️
This study presents an inspecting system using real-time control unmanned aerial vehicles (UAVs) to investigate structural surfaces. The system operates under favourable weather conditions to inspect a target structure, which is the Wentworth light rail base structure in this study. The system includes a drone, a GoPro HERO4 camera, a controller and a mobile phone. The drone takes off the ground manually in the testing field to collect the data requiring for later analysis. The images are taken through HERO 4 camera and then transferred in real time to the remote processing unit such as a ground control station by the wireless connection established by a Wi-Fi router. An image processing method has been proposed to detect defects or damages such as cracks. The method based on intensity histogram algorithms to exploit the pixel group related to the crack contained in the low intensity interval. Experiments, simulation and comparisons have been conducted to evaluate the performance and validity of the proposed system.
8.Self-supervised Transfer Learning for Instance Segmentation through Physical Interaction ⬇️
Instance segmentation of unknown objects from images is regarded as relevant for several robot skills including grasping, tracking and object sorting. Recent results in computer vision have shown that large hand-labeled datasets enable high segmentation performance. To overcome the time-consuming process of manually labeling data for new environments, we present a transfer learning approach for robots that learn to segment objects by interacting with their environment in a self-supervised manner. Our robot pushes unknown objects on a table and uses information from optical flow to create training labels in the form of object masks. To achieve this, we fine-tune an existing DeepMask network for instance segmentation on the self-labeled training data acquired by the robot. We evaluate our trained network (SelfDeepMask) on a set of real images showing challenging and cluttered scenes with novel objects. Here, SelfDeepMask outperforms the DeepMask network trained on the COCO dataset by 9.5% in average precision. Furthermore, we combine our approach with recent approaches for training with noisy labels in order to better cope with induced label noise.
9.MaskFace: multi-task face and landmark detector ⬇️
Currently in the domain of facial analysis single task approaches for face detection and landmark localization dominate. In this paper we draw attention to multi-task models solving both tasks simultaneously. We present a highly accurate model for face and landmark detection. The method, called MaskFace, extends previous face detection approaches by adding a keypoint prediction head. The new keypoint head adopts ideas of Mask R-CNN by extracting facial features with a RoIAlign layer. The keypoint head adds small computational overhead in the case of few faces in the image while improving the accuracy dramatically. We evaluate MaskFace's performance on a face detection task on the AFW, PASCAL face, FDDB, WIDER FACE datasets and a landmark localization task on the AFLW, 300W datasets. For both tasks MaskFace achieves state-of-the-art results outperforming many of single-task and multi-task models.
10.Uncertainty Estimation in Deep 2D Echocardiography Segmentation ⬇️
2D echocardiography is the most common imaging modality for cardiovascular diseases. The portability and relatively low-cost nature of Ultrasound (US) enable the US devices needed for performing echocardiography to be made widely available. However, acquiring and interpreting cardiac US images is operator dependent, limiting its use to only places where experts are present. Recently, Deep Learning (DL) has been used in 2D echocardiography for automated view classification, and structure and function assessment. Although these recent works show promise in developing computer-guided acquisition and automated interpretation of echocardiograms, most of these methods do not model and estimate uncertainty which can be important when testing on data coming from a distribution further away from that of the training data. Uncertainty estimates can be beneficial both during the image acquisition phase (by providing real-time feedback to the operator on acquired image's quality), and during automated measurement and interpretation. The performance of uncertainty models and quantification metric may depend on the prediction task and the models being compared. Hence, to gain insight of uncertainty modelling for left ventricular segmentation from US images, we compare three ensembling based uncertainty models quantified using four different metrics (one newly proposed) on state-of-the-art baseline networks using two publicly available echocardiogram datasets. We further demonstrate how uncertainty estimation can be used to automatically reject poor quality images and improve state-of-the-art segmentation results.
11.Localizing Firearm Carriers by Identifying Human-Object Pairs ⬇️
Visual identification of gunmen in a crowd is a challenging problem, that requires resolving the association of a person with an object (firearm). We present a novel approach to address this problem, by defining human-object interaction (and non-interaction) bounding boxes. In a given image, human and firearms are separately detected. Each detected human is paired with each detected firearm, allowing us to create a paired bounding box that contains both object and the human. A network is trained to classify these paired-bounding-boxes into human carrying the identified firearm or not. Extensive experiments were performed to evaluate effectiveness of the algorithm, including exploiting full pose of the human, hand key-points, and their association with the firearm. The knowledge of spatially localized features is key to success of our method by using multi-size proposals with adaptive average pooling. We have also extended a previously firearm detection dataset, by adding more images and tagging in extended dataset the human-firearm pairs (including bounding boxes for firearms and gunmen). The experimental results (\textit{78.5 $AP_{hold}$}) demonstrate effectiveness of the proposed method.
12.Patch Attack for Automatic Check-out ⬇️
Adversarial examples are inputs with imperceptible perturbations that easily misleading deep neural networks(DNNs). Recently, adversarial patch, with noise confined to a small and localized patch, has emerged for its easy feasibility in real-world scenarios. However, existing strategies failed to generate adversarial patches with strong generalization ability. In other words, the adversarial patches were input-specific and failed to attack images from all classes, especially unseen ones during training. To address the problem, this paper proposes a bias-based framework to generate class-agnostic universal adversarial patches with strong generalization ability, which exploits both the perceptual and semantic bias of models. Regarding the perceptual bias, since DNNs are strongly biased towards textures, we exploit the hard examples which convey strong model uncertainties and extract a textural patch prior from them by adopting the style similarities. The patch prior is more close to decision boundaries and would promote attacks. To further alleviate the heavy dependency on large amounts of data in training universal attacks, we further exploit the semantic bias. As the class-wise preference, prototypes are introduced and pursued by maximizing the multi-class margin to help universal training. Taking AutomaticCheck-out (ACO) as the typical scenario, extensive experiments including white-box and black-box settings in both digital-world(RPC, the largest ACO related dataset) and physical-world scenario(Taobao and JD, the world' s largest online shopping platforms) are conducted. Experimental results demonstrate that our proposed framework outperforms state-of-the-art adversarial patch attack methods.
13.On the Value of Out-of-Distribution Testing: An Example of Goodhart's Law ⬇️
Out-of-distribution (OOD) testing is increasingly popular for evaluating a machine learning system's ability to generalize beyond the biases of a training set. OOD benchmarks are designed to present a different joint distribution of data and labels between training and test time. VQA-CP has become the standard OOD benchmark for visual question answering, but we discovered three troubling practices in its current use. First, most published methods rely on explicit knowledge of the construction of the OOD splits. They often rely on ``inverting'' the distribution of labels, e.g. answering mostly 'yes' when the common training answer is 'no'. Second, the OOD test set is used for model selection. Third, a model's in-domain performance is assessed after retraining it on in-domain splits (VQA v2) that exhibit a more balanced distribution of labels. These three practices defeat the objective of evaluating generalization, and put into question the value of methods specifically designed for this dataset. We show that embarrassingly-simple methods, including one that generates answers at random, surpass the state of the art on some question types. We provide short- and long-term solutions to avoid these pitfalls and realize the benefits of OOD evaluation.
14.An Auto-Context Deformable Registration Network for Infant Brain MRI ⬇️
Deformable image registration is fundamental to longitudinal and population analysis. Geometric alignment of the infant brain MR images is challenging, owing to rapid changes in image appearance in association with brain development. In this paper, we propose an infant-dedicated deep registration network that uses the auto-context strategy to gradually refine the deformation fields to obtain highly accurate correspondences. Instead of training multiple registration networks, our method estimates the deformation fields by invoking a single network multiple times for iterative deformation refinement. The final deformation field is obtained by the incremental composition of the deformation fields. Experimental results in comparison with state-of-the-art registration methods indicate that our method achieves higher accuracy while at the same time preserves the smoothness of the deformation fields. Our implementation is available online.
15.Holistic Parameteric Reconstruction of Building Models from Point Clouds ⬇️
Building models are conventionally reconstructed by building roof points planar segmentation and then using a topology graph to group the planes together. Roof edges and vertices are then mathematically represented by intersecting segmented planes. Technically, such solution is based on sequential local fitting, i.e., the entire data of one building are not simultaneously participating in determining the building model. As a consequence, the solution is lack of topological integrity and geometric rigor. Fundamentally different from this traditional approach, we propose a holistic parametric reconstruction method which means taking into consideration the entire point clouds of one building simultaneously. In our work, building models are reconstructed from predefined parametric (roof) primitives. We first use a well-designed deep neural network to segment and identify primitives in the given building point clouds. A holistic optimization strategy is then introduced to simultaneously determine the parameters of a segmented primitive. In the last step, the optimal parameters are used to generate a watertight building model in CityGML format. The airborne LiDAR dataset RoofN3D with predefined roof types is used for our test. It is shown that PointNet++ applied to the entire dataset can achieve an accuracy of 83% for primitive classification. For a subset of 910 buildings in RoofN3D, the holistic approach is then used to determine the parameters of primitives and reconstruct the buildings. The achieved overall quality of reconstruction is 0.08 meters for point-surface-distance or 0.7 times RMSE of the input LiDAR points. The study demonstrates the efficiency and capability of the proposed approach and its potential to handle large scale urban point clouds.
16.Deep Learning Guided Building Reconstruction from Satellite Imagery-derived Point Clouds ⬇️
3D urban reconstruction of buildings from remotely sensed imagery has drawn significant attention during the past two decades. While aerial imagery and LiDAR provide higher resolution, satellite imagery is cheaper and more efficient to acquire for large scale need. However, the high, orbital altitude of satellite observation brings intrinsic challenges, like unpredictable atmospheric effect, multi view angles, significant radiometric differences due to the necessary multiple views, diverse land covers and urban structures in a scene, small base-height ratio or narrow field of view, all of which may degrade 3D reconstruction quality. To address these major challenges, we present a reliable and effective approach for building model reconstruction from the point clouds generated from multi-view satellite images. We utilize multiple types of primitive shapes to fit the input point cloud. Specifically, a deep-learning approach is adopted to distinguish the shape of building roofs in complex and yet noisy scenes. For points that belong to the same roof shape, a multi-cue, hierarchical RANSAC approach is proposed for efficient and reliable segmenting and reconstructing the building point cloud. Experimental results over four selected urban areas (0.34 to 2.04 sq km in size) demonstrate the proposed method can generate detailed roof structures under noisy data environments. The average successful rate for building shape recognition is 83.0%, while the overall completeness and correctness are over 70% with reference to ground truth created from airborne lidar. As the first effort to address the public need of large scale city model generation, the development is deployed as open source software.
17.Retrieving and Highlighting Action with Spatiotemporal Reference ⬇️
In this paper, we present a framework that jointly retrieves and spatiotemporally highlights actions in videos by enhancing current deep cross-modal retrieval methods. Our work takes on the novel task of action highlighting, which visualizes where and when actions occur in an untrimmed video setting. Action highlighting is a fine-grained task, compared to conventional action recognition tasks which focus on classification or window-based localization. Leveraging weak supervision from annotated captions, our framework acquires spatiotemporal relevance maps and generates local embeddings which relate to the nouns and verbs in captions. Through experiments, we show that our model generates various maps conditioned on different actions, in which conventional visual reasoning methods only go as far as to show a single deterministic saliency map. Also, our model improves retrieval recall over our baseline without alignment by 2-3% on the MSR-VTT dataset.
18.MOTS: Multiple Object Tracking for General Categories Based On Few-Shot Method ⬇️
Most modern Multi-Object Tracking (MOT) systems typically apply REID-based paradigm to hold a balance between computational efficiency and performance. In the past few years, numerous attempts have been made to perfect the systems. Although they presented favorable performance, they were constrained to track specified category. Drawing on the ideas of few shot method, we pioneered a new multi-target tracking system, named MOTS, which is based on metrics but not limited to track specific category. It contains two stages in series: In the first stage, we design the self-Adaptive-matching module to perform simple targets matching, which can complete 88.76% assignments without sacrificing performance on MOT16 training set. In the second stage, a Fine-match Network was carefully designed for unmatched targets. With a newly built TRACK-REID data-set, the Fine-match Network can perform matching of 31 category targets, even generalizes to unseen categories.
19.Learning from a Lightweight Teacher for Efficient Knowledge Distillation ⬇️
Knowledge Distillation (KD) is an effective framework for compressing deep learning models, realized by a student-teacher paradigm requiring small student networks to mimic the soft target generated by well-trained teachers. However, the teachers are commonly assumed to be complex and need to be trained on the same datasets as students. This leads to a time-consuming training process. The recent study shows vanilla KD plays a similar role as label smoothing and develops teacher-free KD, being efficient and mitigating the issue of learning from heavy teachers. But because teacher-free KD relies on manually-crafted output distributions kept the same for all data instances belonging to the same class, its flexibility and performance are relatively limited. To address the above issues, this paper proposes en efficient knowledge distillation learning framework LW-KD, short for lightweight knowledge distillation. It firstly trains a lightweight teacher network on a synthesized simple dataset, with an adjustable class number equal to that of a target dataset. The teacher then generates soft target whereby an enhanced KD loss could guide student learning, which is a combination of KD loss and adversarial loss for making student output indistinguishable from the output of the teacher. Experiments on several public datasets with different modalities demonstrate LWKD is effective and efficient, showing the rationality of its main design principles.
20.Adversarial Attacks for Embodied Agents ⬇️
Adversarial attacks are valuable for providing insights into the blind-spots of deep learning models and help improve their robustness. Existing work on adversarial attacks have mainly focused on static scenes; however, it remains unclear whether such attacks are effective against embodied agents, which could navigate and interact with a dynamic environment. In this work, we take the first step to study adversarial attacks for embodied agents. In particular, we generate spatiotemporal perturbations to form 3D adversarial examples, which exploit the interaction history in both the temporal and spatial dimensions. Regarding the temporal dimension, since agents make predictions based on historical observations, we develop a trajectory attention module to explore scene view contributions, which further help localize 3D objects appeared with the highest stimuli. By conciliating with clues from the temporal dimension, along the spatial dimension, we adversarially perturb the physical properties (e.g., texture and 3D shape) of the contextual objects that appeared in the most important scene views. Extensive experiments on the EQA-v1 dataset for several embodied tasks in both the white-box and black-box settings have been conducted, which demonstrate that our perturbations have strong attack and generalization abilities.
21.Sketch-BERT: Learning Sketch Bidirectional Encoder Representation from Transformers by Self-supervised Learning of Sketch Gestalt ⬇️
Previous researches of sketches often considered sketches in pixel format and leveraged CNN based models in the sketch understanding. Fundamentally, a sketch is stored as a sequence of data points, a vector format representation, rather than the photo-realistic image of pixels. SketchRNN studied a generative neural representation for sketches of vector format by Long Short Term Memory networks (LSTM). Unfortunately, the representation learned by SketchRNN is primarily for the generation tasks, rather than the other tasks of recognition and retrieval of sketches. To this end and inspired by the recent BERT model, we present a model of learning Sketch Bidirectional Encoder Representation from Transformer (Sketch-BERT). We generalize BERT to sketch domain, with the novel proposed components and pre-training algorithms, including the newly designed sketch embedding networks, and the self-supervised learning of sketch gestalt. Particularly, towards the pre-training task, we present a novel Sketch Gestalt Model (SGM) to help train the Sketch-BERT. Experimentally, we show that the learned representation of Sketch-BERT can help and improve the performance of the downstream tasks of sketch recognition, sketch retrieval, and sketch gestalt.
22.Associating Multi-Scale Receptive Fields for Fine-grained Recognition ⬇️
Extracting and fusing part features have become the key of fined-grained image recognition. Recently, Non-local (NL) module has shown excellent improvement in image recognition. However, it lacks the mechanism to model the interactions between multi-scale part features, which is vital for fine-grained recognition. In this paper, we propose a novel cross-layer non-local (CNL) module to associate multi-scale receptive fields by two operations. First, CNL computes correlations between features of a query layer and all response layers. Second, all response features are weighted according to the correlations and are added to the query features. Due to the interactions of cross-layer features, our model builds spatial dependencies among multi-level layers and learns more discriminative features. In addition, we can reduce the aggregation cost if we set low-dimensional deep layer as query layer. Experiments are conducted to show our model achieves or surpasses state-of-the-art results on three benchmark datasets of fine-grained classification. Our codes can be found at this http URL.
23.Increasing-Margin Adversarial (IMA) Training to Improve Adversarial Robustness of Neural Networks ⬇️
Convolutional neural network (CNN) has surpassed traditional methods for med-ical image classification. However, CNN is vulnerable to adversarial attacks which may lead to disastrous consequences in medical applications. Although adversarial noises are usually generated by attack algorithms, white-noise-induced adversarial samples can exist, and therefore the threats are real. In this study, we propose a novel training method, named IMA, to improve the robust-ness of CNN against adversarial noises. During training, the IMA method in-creases the margins of training samples in the input space, i.e., moving CNN de-cision boundaries far away from the training samples to improve robustness. The IMA method is evaluated on four publicly available datasets under strong 100-PGD white-box adversarial attacks, and the results show that the proposed meth-od significantly improved CNN classification accuracy on noisy data while keep-ing a relatively high accuracy on clean data. We hope our approach may facilitate the development of robust applications in medical field.
24.Domain Adaptive Relational Reasoning for 3D Multi-Organ Segmentation ⬇️
In this paper, we present a novel unsupervised domain adaptation (UDA) method, named Domain Adaptive Relational Reasoning (DARR), to generalize 3D multi-organ segmentation models to medical data collected from different scanners and/or protocols (domains). Our method is inspired by the fact that the spatial relationship between internal structures in medical images is relatively fixed, e.g., a spleen is always located at the tail of a pancreas, which serves as a latent variable to transfer the knowledge shared across multiple domains. We formulate the spatial relationship by solving a jigsaw puzzle task, i.e., recovering a CT scan from its shuffled patches, and jointly train it with the organ segmentation task. To guarantee the transferability of the learned spatial relationship to multiple domains, we additionally introduce two schemes: 1) Employing a super-resolution network also jointly trained with the segmentation model to standardize medical images from different domain to a certain spatial resolution; 2) Adapting the spatial relationship for a test image by test-time jigsaw puzzle training. Experimental results show that our method improves the performance by 29.60% DSC on target datasets on average without using any data from the target domain during training.
25.Two-View Fine-grained Classification of Plant Species ⬇️
Automatic plant classification is a challenging problem due to the wide biodiversity of the existing plant species in a fine-grained scenario. Powerful deep learning architectures have been used to improve the classification performance in such a fine-grained problem, but usually building models that are highly dependent on a large training dataset and which are not scalable. In this paper, we propose a novel method based on a two-view leaf image representation and a hierarchical classification strategy for fine-grained recognition of plant species. It uses the botanical taxonomy as a basis for a coarse-to-fine strategy applied to identify the plant genus and species. The two-view representation provides complementary global and local features of leaf images. A deep metric based on Siamese convolutional neural networks is used to reduce the dependence on a large number of training samples and make the method scalable to new plant species. The experimental results on two challenging fine-grained datasets of leaf images (i.e. LifeCLEF 2015 and LeafSnap) have shown the effectiveness of the proposed method, which achieved recognition accuracy of 0.87 and 0.96 respectively.
26.Cross-filter compression for CNN inference acceleration ⬇️
Convolution neural network demonstrates great capability for multiple tasks, such as image classification and many others. However, much resource is required to train a network. Hence much effort has been made to accelerate neural network by reducing precision of weights, activation, and gradient. However, these filter-wise quantification methods exist a natural upper limit, caused by the size of the kernel. Meanwhile, with the popularity of small kernel, the natural limit further decrease. To address this issue, we propose a new cross-filter compression method that can provide
$\sim32\times$ memory savings and$122\times$ speed up in convolution operations. In our method, all convolution filters are quantized to given bits and spatially adjacent filters share the same scaling factor. Our compression method, based on Binary-Weight and XNOR-Net separately, is evaluated on CIFAR-10 and ImageNet dataset with widely used network structures, such as ResNet and VGG, and witness tolerable accuracy loss compared to state-of-the-art quantification methods.
27.Efficient Image Gallery Representations at Scale Through Multi-Task Learning ⬇️
Image galleries provide a rich source of diverse information about a product which can be leveraged across many recommendation and retrieval applications. We study the problem of building a universal image gallery encoder through multi-task learning (MTL) approach and demonstrate that it is indeed a practical way to achieve generalizability of learned representations to new downstream tasks. Additionally, we analyze the relative predictive performance of MTL-trained solutions against optimal and substantially more expensive solutions, and find signals that MTL can be a useful mechanism to address sparsity in low-resource binary tasks.
28.A Novel Technique Combining Image Processing, Plant Development Properties, and the Hungarian Algorithm, to Improve Leaf Detection in Maize ⬇️
Manual determination of plant phenotypic properties such as plant architecture, growth, and health is very time consuming and sometimes destructive. Automatic image analysis has become a popular approach. This research aims to identify the position (and number) of leaves from a temporal sequence of high-quality indoor images consisting of multiple views, focussing in particular of images of maize. The procedure used a segmentation on the images, using the convex hull to pick the best view at each time step, followed by a skeletonization of the corresponding image. To remove skeleton spurs, a discrete skeleton evolution pruning process was applied. Pre-existing statistics regarding maize development was incorporated to help differentiate between true leaves and false leaves. Furthermore, for each time step, leaves were matched to those of the previous and next three days using the graph-theoretic Hungarian algorithm. This matching algorithm can be used to both remove false positives, and also to predict true leaves, even if they were completely occluded from the image itself. The algorithm was evaluated using an open dataset consisting of 13 maize plants across 27 days from two different views. The total number of true leaves from the dataset was 1843, and our proposed techniques detect a total of 1690 leaves including 1674 true leaves, and only 16 false leaves, giving a recall of 90.8%, and a precision of 99.0%.
29.Patch based Colour Transfer using SIFT Flow ⬇️
We propose a new colour transfer method with Optimal Transport (OT) to transfer the colour of a sourceimage to match the colour of a target image of the same scene that may exhibit large motion changes betweenimages. By definition OT does not take into account any available information about correspondences whencomputing the optimal solution. To tackle this problem we propose to encode overlapping neighborhoodsof pixels using both their colour and spatial correspondences estimated using motion estimation. We solvethe high dimensional problem in 1D space using an iterative projection approach. We further introducesmoothing as part of the iterative algorithms for solving optimal transport namely Iterative DistributionTransport (IDT) and its variant the Sliced Wasserstein Distance (SWD). Experiments show quantitative andqualitative improvements over previous state of the art colour transfer methods.
30.U$^2$-Net: Going Deeper with Nested U-Structure for Salient Object Detection ⬇️
In this paper, we design a simple yet powerful deep network architecture, U$^2$-Net, for salient object detection (SOD). The architecture of our U$^2$-Net is a two-level nested U-structure. The design has the following advantages: (1) it is able to capture more contextual information from different scales thanks to the mixture of receptive fields of different sizes in our proposed ReSidual U-blocks (RSU), (2) it increases the depth of the whole architecture without significantly increasing the computational cost because of the pooling operations used in these RSU blocks. This architecture enables us to train a deep network from scratch without using backbones from image classification tasks. We instantiate two models of the proposed architecture, U$^2$-Net (176.3 MB, 30 FPS on GTX 1080Ti GPU) and U$^2$-Net$^{\dagger}$ (4.7 MB, 40 FPS), to facilitate the usage in different environments. Both models achieve competitive performance on six SOD datasets. The code is available: this https URL.
31.Identifying Statistical Bias in Dataset Replication ⬇️
Dataset replication is a useful tool for assessing whether improvements in test accuracy on a specific benchmark correspond to improvements in models' ability to generalize reliably. In this work, we present unintuitive yet significant ways in which standard approaches to dataset replication introduce statistical bias, skewing the resulting observations. We study ImageNet-v2, a replication of the ImageNet dataset on which models exhibit a significant (11-14%) drop in accuracy, even after controlling for a standard human-in-the-loop measure of data quality. We show that after correcting for the identified statistical bias, only an estimated
$3.6% \pm 1.5%$ of the original$11.7% \pm 1.0%$ accuracy drop remains unaccounted for. We conclude with concrete recommendations for recognizing and avoiding bias in dataset replication. Code for our study is publicly available at this http URL .
32.Advances in Computer Vision in Gastric Cancer: Potential Efficient Tools for Diagnosis ⬇️
Early and rapid diagnosis of gastric cancer is a great challenge for clinical doctors. Dramatic progress of computer vision on gastric cancer has been made recently and this review focused on advances during the past five years. Different methods for data generation and augmentation have been presented, and various approaches to extract discriminative features compared and evaluated. Classification and segmentation techniques are carefully discussed for assisting more precise diagnosis and timely treatment. Application of those methods will greatly reduce the labor and time consumed for the diagnosis of gastric cancers.
33.The Skincare project, an interactive deep learning system for differential diagnosis of malignant skin lesions. Technical Report ⬇️
A shortage of dermatologists causes long wait times for patients who seek dermatologic care. In addition, the diagnostic accuracy of general practitioners has been reported to be lower than the accuracy of artificial intelligence software. This article describes the Skincare project (H2020, EIT Digital). Contributions include enabling technology for clinical decision support based on interactive machine learning (IML), a reference architecture towards a Digital European Healthcare Infrastructure (also cf. EIT MCPS), technical components for aggregating digitised patient information, and the integration of decision support technology into clinical test-bed environments. However, the main contribution is a diagnostic and decision support system in dermatology for patients and doctors, an interactive deep learning system for differential diagnosis of malignant skin lesions. In this article, we describe its functionalities and the user interfaces to facilitate machine learning from human input. The baseline deep learning system, which delivers state-of-the-art results and the potential to augment general practitioners and even dermatologists, was developed and validated using de-identified cases from a dermatology image data base (ISIC), which has about 20000 cases for development and validation, provided by board-certified dermatologists defining the reference standard for every case. ISIC allows for differential diagnosis, a ranked list of eight diagnoses, that is used to plan treatments in the common setting of diagnostic ambiguity. We give an overall description of the outcome of the Skincare project, and we focus on the steps to support communication and coordination between humans and machine in IML. This is an integral part of the development of future cognitive assistants in the medical domain, and we describe the necessary intelligent user interfaces.
34.hidden markov random fields and cuckoo search method for medical image segmentation ⬇️
Segmentation of medical images is an essential part in the process of diagnostics. Physicians require an automatic, robust and valid results. Hidden Markov Random Fields (HMRF) provide powerful model. This latter models the segmentation problem as the minimization of an energy function. Cuckoo search (CS) algorithm is one of the recent nature-inspired meta-heuristic algorithms. It has shown its efficiency in many engineering optimization problems. In this paper, we use three cuckoo search algorithm to achieve medical image segmentation.
35.Learning to segment clustered amoeboid cells from brightfield microscopy via multi-task learning with adaptive weight selection ⬇️
Detecting and segmenting individual cells from microscopy images is critical to various life science applications. Traditional cell segmentation tools are often ill-suited for applications in brightfield microscopy due to poor contrast and intensity heterogeneity, and only a small subset are applicable to segment cells in a cluster. In this regard, we introduce a novel supervised technique for cell segmentation in a multi-task learning paradigm. A combination of a multi-task loss, based on the region and cell boundary detection, is employed for an improved prediction efficiency of the network. The learning problem is posed in a novel min-max framework which enables adaptive estimation of the hyper-parameters in an automatic fashion. The region and cell boundary predictions are combined via morphological operations and active contour model to segment individual cells.
The proposed methodology is particularly suited to segment touching cells from brightfield microscopy images without manual interventions. Quantitatively, we observe an overall Dice score of 0.93 on the validation set, which is an improvement of over 15.9% on a recent unsupervised method, and outperforms the popular supervised U-net algorithm by at least$5.8%$ on average.
36.AdaptiveWeighted Attention Network with Camera Spectral Sensitivity Prior for Spectral Reconstruction from RGB Images ⬇️
Recent promising effort for spectral reconstruction (SR) focuses on learning a complicated mapping through using a deeper and wider convolutional neural networks (CNNs). Nevertheless, most CNN-based SR algorithms neglect to explore the camera spectral sensitivity (CSS) prior and interdependencies among intermediate features, thus limiting the representation ability of the network and performance of SR. To conquer these issues, we propose a novel adaptive weighted attention network (AWAN) for SR, whose backbone is stacked with multiple dual residual attention blocks (DRAB) decorating with long and short skip connections to form the dual residual learning. Concretely, we investigate an adaptive weighted channel attention (AWCA) module to reallocate channel-wise feature responses via integrating correlations between channels. Furthermore, a patch-level second-order non-local (PSNL) module is developed to capture long-range spatial contextual information by second-order non-local operations for more powerful feature representations. Based on the fact that the recovered RGB images can be projected by the reconstructed hyperspectral image (HSI) and the given CSS function, we incorporate the discrepancies of the RGB images and HSIs as a finer constraint for more accurate reconstruction. Experimental results demonstrate the effectiveness of our proposed AWAN network in terms of quantitative comparison and perceptual quality over other state-of-the-art SR methods. In the NTIRE 2020 Spectral Reconstruction Challenge, our entries obtain the 1st ranking on the Clean track and the 3rd place on the Real World track. Codes are available at this https URL.
37.Synthesizing Unrestricted False Positive Adversarial Objects Using Generative Models ⬇️
Adversarial examples are data points misclassified by neural networks. Originally, adversarial examples were limited to adding small perturbations to a given image. Recent work introduced the generalized concept of unrestricted adversarial examples, without limits on the added perturbations. In this paper, we introduce a new category of attacks that create unrestricted adversarial examples for object detection. Our key idea is to generate adversarial objects that are unrelated to the classes identified by the target object detector. Different from previous attacks, we use off-the-shelf Generative Adversarial Networks (GAN), without requiring any further training or modification. Our method consists of searching over the latent normal space of the GAN for adversarial objects that are wrongly identified by the target object detector. We evaluate this method on the commonly used Faster R-CNN ResNet-101, Inception v2 and SSD Mobilenet v1 object detectors using logo generative iWGAN-LC and SNGAN trained on CIFAR-10. The empirical results show that the generated adversarial objects are indistinguishable from non-adversarial objects generated by the GANs, transferable between the object detectors and robust in the physical world. This is the first work to study unrestricted false positive adversarial examples for object detection.
38.Assertion Detection in Multi-Label Clinical Text using Scope Localization ⬇️
Multi-label sentences (text) in the clinical domain result from the rich description of scenarios during patient care. The state-of-theart methods for assertion detection mostly address this task in the setting of a single assertion label per sentence (text). In addition, few rules based and deep learning methods perform negation/assertion scope detection on single-label text. It is a significant challenge extending these methods to address multi-label sentences without diminishing performance. Therefore, we developed a convolutional neural network (CNN) architecture to localize multiple labels and their scopes in a single stage end-to-end fashion, and demonstrate that our model performs atleast 12% better than the state-of-the-art on multi-label clinical text.
39.Structural Residual Learning for Single Image Rain Removal ⬇️
To alleviate the adverse effect of rain streaks in image processing tasks, CNN-based single image rain removal methods have been recently proposed. However, the performance of these deep learning methods largely relies on the covering range of rain shapes contained in the pre-collected training rainy-clean image pairs. This makes them easily trapped into the overfitting-to-the-training-samples issue and cannot finely generalize to practical rainy images with complex and diverse rain streaks. Against this generalization issue, this study proposes a new network architecture by enforcing the output residual of the network possess intrinsic rain structures. Such a structural residual setting guarantees the rain layer extracted by the network finely comply with the prior knowledge of general rain streaks, and thus regulates sound rain shapes capable of being well extracted from rainy images in both training and predicting stages. Such a general regularization function naturally leads to both its better training accuracy and testing generalization capability even for those non-seen rain configurations. Such superiority is comprehensively substantiated by experiments implemented on synthetic and real datasets both visually and quantitatively as compared with current state-of-the-art methods.
40.A Self-ensembling Framework for Semi-supervised Knee Osteoarthritis Localization and Classification with Dual-Consistency ⬇️
Knee osteoarthritis (OA) is one of the most common musculoskeletal disorders and requires early-stage diagnosis. Nowadays, the deep convolutional neural networks have achieved greatly in the computer-aided diagnosis field. However, the construction of the deep learning models usually requires great amounts of annotated data, which is generally high-cost. In this paper, we propose a novel approach for knee OA diagnosis, including severity classification and lesion localization. Particularly, we design a self-ensembling framework, which is composed of a student network and a teacher network with the same structure. The student network learns from both labeled data and unlabeled data and the teacher network averages the student model weights through the training course. A novel attention loss function is developed to obtain accurate attention masks. With dual-consistency checking of the attention in the lesion classification and localization, the two networks can gradually optimize the attention distribution and improve the performance of each other, whereas the training relies on partially labeled data only and follows the semi-supervised manner. Experiments show that the proposed method can significantly improve the self-ensembling performance in both knee OA classification and localization, and also greatly reduce the needs of annotated data.
41.Regularization Methods for Generative Adversarial Networks: An Overview of Recent Studies ⬇️
Despite its short history, Generative Adversarial Network (GAN) has been extensively studied and used for various tasks, including its original purpose, i.e., synthetic sample generation. However, applying GAN to different data types with diverse neural network architectures has been hindered by its limitation in training, where the model easily diverges. Such a notorious training of GANs is well known and has been addressed in numerous studies. Consequently, in order to make the training of GAN stable, numerous regularization methods have been proposed in recent years. This paper reviews the regularization methods that have been recently introduced, most of which have been published in the last three years. Specifically, we focus on general methods that can be commonly used regardless of neural network architectures. To explore the latest research trends in the regularization for GANs, the methods are classified into several groups by their operation principles, and the differences between the methods are analyzed. Furthermore, to provide practical knowledge of using these methods, we investigate popular methods that have been frequently employed in state-of-the-art GANs. In addition, we discuss the limitations in existing methods and propose future research directions.
42.A New Validity Index for Fuzzy-Possibilistic C-Means Clustering ⬇️
In some complicated datasets, due to the presence of noisy data points and outliers, cluster validity indices can give conflicting results in determining the optimal number of clusters. This paper presents a new validity index for fuzzy-possibilistic c-means clustering called Fuzzy-Possibilistic (FP) index, which works well in the presence of clusters that vary in shape and density. Moreover, FPCM like most of the clustering algorithms is susceptible to some initial parameters. In this regard, in addition to the number of clusters, FPCM requires a priori selection of the degree of fuzziness and the degree of typicality. Therefore, we presented an efficient procedure for determining their optimal values. The proposed approach has been evaluated using several synthetic and real-world datasets. Final computational results demonstrate the capabilities and reliability of the proposed approach compared with several well-known fuzzy validity indices in the literature. Furthermore, to clarify the ability of the proposed method in real applications, the proposed method is implemented in microarray gene expression data clustering and medical image segmentation.
43.Improve robustness of DNN for ECG signal classification:a noise-to-signal ratio perspective ⬇️
Electrocardiogram (ECG) is the most widely used diagnostic tool to monitor the condition of the cardiovascular system. Deep neural networks (DNNs), have been developed in many research labs for automatic interpretation of ECG signals to identify potential abnormalities in patient hearts. Studies have shown that given a sufficiently large amount of data, the classification accuracy of DNNs could reach human-expert cardiologist level. A DNN-based automated ECG diagnostic system would be an affordable solution for patients in developing countries where human-expert cardiologist are lacking. However, despite of the excellent performance in classification accuracy, it has been shown that DNNs are highly vulnerable to adversarial attacks: subtle changes in input of a DNN can lead to a wrong classification output with high confidence. Thus, it is challenging and essential to improve adversarial robustness of DNNs for ECG signal classification, a life-critical application. In this work, we proposed to improve DNN robustness from the perspective of noise-to-signal ratio (NSR) and developed two methods to minimize NSR during training process. We evaluated the proposed methods on PhysionNets MIT-BIH dataset, and the results show that our proposed methods lead to an enhancement in robustness against PGD adversarial attack and SPSA attack, with a minimal change in accuracy on clean data.
44.An Artificial-intelligence/Statistics Solution to Quantify Material Distortion for Thermal Compensation in Additive Manufacturing ⬇️
In this paper, we introduce a probabilistic statistics solution or artificial intelligence (AI) approach to identify and quantify permanent (non-zero strain) continuum/material deformation only based on the scanned material data in the spatial configuration and the shape of the initial design configuration or the material configuration. The challenge of this problem is that we only know the scanned material data in the spatial configuration and the shape of the design configuration of three-dimensional (3D) printed products, whereas for a specific scanned material point we do not know its corresponding material coordinates in the initial or designed referential configuration, provided that we do not know the detailed information on actual physical deformation process. Different from physics-based modeling, the method developed here is a data-driven artificial intelligence method, which solves the problem with incomplete deformation data or with missing information of actual physical deformation process. We coined the method is an AI-based material deformation finding algorithm.
This method has practical significance and important applications in finding and designing thermal compensation configuration of a 3D printed product in additive manufacturing, which is at the heart of the cutting edge 3D printing technology. In this paper, we demonstrate that the proposed AI continuum/material deformation finding approach can accurately find permanent thermal deformation configuration for a complex 3D printed structure component, and hence to identify the thermal compensation design configuration in order to minimizing the impact of temperature fluctuations on 3D printed structure components that are sensitive to changes of temperature.
45.Translating Video Recordings of Mobile App Usages into Replayable Scenarios ⬇️
Screen recordings of mobile applications are easy to obtain and capture a wealth of information pertinent to software developers (e.g., bugs or feature requests), making them a popular mechanism for crowdsourced app feedback. Thus, these videos are becoming a common artifact that developers must manage. In light of unique mobile development constraints, including swift release cycles and rapidly evolving platforms, automated techniques for analyzing all types of rich software artifacts provide benefit to mobile developers. Unfortunately, automatically analyzing screen recordings presents serious challenges, due to their graphical nature, compared to other types of (textual) artifacts. To address these challenges, this paper introduces V2S, a lightweight, automated approach for translating video recordings of Android app usages into replayable scenarios. V2S is based primarily on computer vision techniques and adapts recent solutions for object detection and image classification to detect and classify user actions captured in a video, and convert these into a replayable test scenario. We performed an extensive evaluation of V2S involving 175 videos depicting 3,534 GUI-based actions collected from users exercising features and reproducing bugs from over 80 popular Android apps. Our results illustrate that V2S can accurately replay scenarios from screen recordings, and is capable of reproducing
$\approx$ 89% of our collected videos with minimal overhead. A case study with three industrial partners illustrates the potential usefulness of V2S from the viewpoint of developers.
46.On the effectiveness of GAN generated cardiac MRIs for segmentation ⬇️
In this work, we propose a Variational Autoencoder (VAE) - Generative Adversarial Networks (GAN) model that can produce highly realistic MRI together with its pixel accurate groundtruth for the application of cine-MR image cardiac segmentation. On one side of our model is a Variational Autoencoder (VAE) trained to learn the latent representations of cardiac shapes. On the other side is a GAN that uses "SPatially-Adaptive (DE)Normalization" (SPADE) modules to generate realistic MR images tailored to a given anatomical map. At test time, the sampling of the VAE latent space allows to generate an arbitrary large number of cardiac shapes, which are fed to the GAN that subsequently generates MR images whose cardiac structure fits that of the cardiac shapes. In other words, our system can generate a large volume of realistic yet labeled cardiac MR images. We show that segmentation with CNNs trained with our synthetic annotated images gets competitive results compared to traditional techniques. We also show that combining data augmentation with our GAN-generated images lead to an improvement in the Dice score of up to 12 percent while allowing for better generalization capabilities on other datasets.
47.Saving the Sonorine: Audio Recovery Using Image Processing and Computer Vision ⬇️
This paper presents a novel technique to recover audio from sonorines, an early 20th century form of analogue sound storage. Our method uses high resolution photographs of sonorines under different lighting conditions to observe the change in reflection behavior of the physical surface features and create a three-dimensional height map of the surface. Sound can then be extracted using height information within the surface's grooves, mimicking a physical stylus on a phonograph. Unlike traditional playback methods, our method has the advantage of being contactless: the medium will not incur damage and wear from being played repeatedly. We compare the results of our technique to a previously successful contactless method using flatbed scans of the sonorines, and conclude with future research that can be applied to this photovisual approach to audio recovery.