What is this?
A curated list of resources for semantic segmentation: network architectures, papers, code implementations, datasets, benchmarks, annotation tools, and losses used in the field.
A curated list of resources for semantic segmentation: network architectures, papers, code implementations, datasets, benchmarks, annotation tools, and losses used in the field.Is this a library or just a list of links?It is purely a curated list (an awesome-list), not a runnable library or framework. Each entry links to a paper and one or more code implementations in frameworks like PyTorch, TensorFlow, Caffe, or Keras.
It is purely a curated list (an awesome-list), not a runnable library or framework. Each entry links to a paper and one or more code implementations in frameworks like PyTorch, TensorFlow, Caffe, or Keras.How is the list organized?It is organized by topic: networks by architecture (semantic, instance-aware, weakly-supervised), RNN-based and GAN-based methods, graphical models, datasets, benchmarks, evaluation code, annotation tools, metrics, and losses, plus specialized sections for medical imaging, satellite images, video segmentation, and autonomous driving.
It is organized by topic: networks by architecture (semantic, instance-aware, weakly-supervised), RNN-based and GAN-based methods, graphical models, datasets, benchmarks, evaluation code, annotation tools, metrics, and losses, plus specialized sections for medical imaging, satellite images, video segmentation, and autonomous driving.Are entries vetted and kept up to date?The list spans entries from 2015 through 2020 and is community maintained with no explicit vetting criteria stated; some sections (like 'Networks by framework') are explicitly marked as an older, less current list.
The list spans entries from 2015 through 2020 and is community maintained with no explicit vetting criteria stated; some sections (like 'Networks by framework') are explicitly marked as an older, less current list.Where should a beginner start?A newcomer can start with classic architectures listed first, like U-Net, SegNet, FCN, and DeepLab, each linking to original papers and multiple framework implementations.
A newcomer can start with classic architectures listed first, like U-Net, SegNet, FCN, and DeepLab, each linking to original papers and multiple framework implementations.What datasets are included for semantic segmentation benchmarking?It lists major datasets such as Cityscapes, Microsoft COCO, ADE20K, PASCAL Context, CamVid, Mapillary Vistas, and KITTI, among others.
It lists major datasets such as Cityscapes, Microsoft COCO, ADE20K, PASCAL Context, CamVid, Mapillary Vistas, and KITTI, among others.Does the list cover instance segmentation too?Yes, it has a dedicated 'Instance aware segmentation' section covering methods like Mask-RCNN, YOLACT, SOLO, and PANet.
Yes, it has a dedicated 'Instance aware segmentation' section covering methods like Mask-RCNN, YOLACT, SOLO, and PANet.Are there specialized sections beyond general computer vision?Yes, it includes dedicated sections for medical image segmentation, satellite image segmentation, video segmentation, and autonomous driving applications.
Yes, it includes dedicated sections for medical image segmentation, satellite image segmentation, video segmentation, and autonomous driving applications.What kind of code links are provided for each method?Each network entry links to one or more open-source implementations in frameworks such as PyTorch, TensorFlow, Caffe, Keras, Torch, MXNet, or Chainer.
Each network entry links to one or more open-source implementations in frameworks such as PyTorch, TensorFlow, Caffe, Keras, Torch, MXNet, or Chainer.What are similar projects?DeepLearning, techniques, CVPR2026-Papers-with-Code, VisionHOPE, awesome-deep-vision, nnUNet_att_position_correction, deep_learning_object_detection, CVPR2024-Paper-Code-Interpretation
DeepLearning, techniques, CVPR2026-Papers-with-Code, VisionHOPE, awesome-deep-vision, nnUNet_att_position_correction, deep_learning_object_detection, CVPR2024-Paper-Code-InterpretationWhich topics is it in?autonomous-driving, medical-imaging, semantic-segmentation, image-segmentation, deeplearning, image-classification, object-detection, papers, tensorflow, evaluation, neural-network, computer-vision
autonomous-driving, medical-imaging, semantic-segmentation, image-segmentation, deeplearning, image-classification, object-detection, papers, tensorflow, evaluation, neural-network, computer-vision