Instance Semantic Segmentation List

This repository contains lists of state-or-art instance semantic segmentation works. Papers and resources are listed below according to method types.

• Paper list
  • detection based
  • segmentation based

Brief introduction

Instance semantic segmentation is a area closely related to detection and semantic segmentation. In particular, it could be seen as detection plus foreground mask. But mostly it is not able to segment non-object pixels such as sky, land etc(which considered as a scene parsing task under semantic segmentation). For quick review related topics, see these survey papers:

Speed/accuracy trade-offs for modern convolutional object detectors, CVPR 2017

Survey of recent progress in semantic image segmentation with CNNs, until 201706

Dataset and benchmark here

Dataset	Train	Val	Link	Note
Pascal VOC 12 Aug	10582	1449	SegVOC12	origin train 1464+SDB
Pascal VOC SDB val	5623	5732	SDB	similar to VOC12 Main
COCO	115k	5k	COCO	coco_2014_minival
CityScapes	5000	/	CityScapes	evaluation server

Note: Pascal VOC could refer to different split of dataset. Original VOC12 segmentation task consists of train/val/test 1464/1449/1456 images respectively without instance information. It is designed for semantic segmentation. SDB provides instance-aware annotations for images from Pascal VOC12. And their split(8k/3k) differ from VOC, so another split from them Pasval VOC SDB val is provided, which is similar to Pascal VOC Main split(5717/5823).

1.Detection-based methods

• PAnet:Path Aggregation Network for Instance Segmentation, CVPR 2018

• MaskLab: Instance Segmentation by Refining Object Detection with Semantic and Direction Features, CVPR 2018

• Mask R-CNN, ICCV 2017 [Detectron][Code(TF)][Code(MXNET)]

• BlitzNet: A Real-Time Deep Network for Scene Understanding, ICCV 2017 [web][code]

• FCIS: Fully Convolutional Instance-Aware Semantic Segmentation, CVPR 2017 [code]

• FastMask: Segment Multi-scale Object Candidates in One Shot, CVPR 2017 [code]

• instance sensitive fully convolutional networks, ECCV 2016

• MNC: Instance-aware Semantic Segmentation via Multi-task Network Cascades, CVPR 2016 [code]

• SharpMask: Learning to Refine Object Segments, ECCV 2016

• DeepMask: Learning to Segment Object Candidates, NIPS 2015 [code]

2.Segmentation-based methods

• SGN: Sequential Grouping Networks for Instance Segmentation, ICCV 2017

• InstanceCut: from Edges to Instances with MultiCut,CVPR 2017

• Pixelwise Instance Segmentation with a Dynamically Instantiated Network, CVPR 2017

• Deep Watershed Transform for Instance Segmentation, CVPR 2017 [code]

• Multi-scale Patch Aggregation (MPA) for Simultaneous Detection and Segmentation, CVPR 2016

3.Others

• End-to-End Instance Segmentation with Recurrent Attention, CVPR 2017

• Recurrent Instance Segmentation, ECCV 2016 [web]

• SDS:Simultaneous Detection and Segmentation, ECCV 2014 [web] [code]

• Hypercolumns for Object Segmentation and Fine-grained Localization , CVPR 2015