Awesome Weak-Shot Learning

In weak-shot learning, all categories are split into non-overlapped base categories and novel categories, in which base categories have full annotations while novel categories only have weak annotations. In different tasks, weak annotation could be provided in different forms, e.g., noisy label for classification, image label for object detection, image label/bounding box for segmentation.

The comparison between weak-shot learning and zero/few-shot learning is illustrated below. In all three settings, all categories are split into non-overlapped base categories and novel categories. In all three settings, base categories have abundant fully-annotated training samples. In zero-shot learning, novel categories have no training samples, so class-level representations are required to bridge the gap between base categories and novel categories. In few-shot learning, novel categories have limited training samples. In weak-shot leanring, novel categories have abundant weakly-annotated training samples.

Contributing

Contributions are welcome. If you wish to contribute, feel free to send a pull request. If you have suggestions for new sections to be included, please raise an issue and discuss before sending a pull request.

Table of Contents

• Survey
• Weak-Shot Classification
• Weak-Shot Object Detection
• Weak-Shot Semantic Segmentation
• Weak-Shot Instance Segmentation

Survey

• Li Niu: "Weak Novel Categories without Tears: A Survey on Weak-Shot Learning." arXiv preprint arXiv:2110.02651 (2021). [arXiv]

Weak-Shot Classification

Base category: clean label; Novel category: noisy label (weak-shot)

• Junjie Chen, Li Niu, Liu Liu, Liqing Zhang: "Weak-shot Fine-grained Classification via Similarity Transfer." NeurIPS (2021) [arXiv] [code]

Weak-Shot Object Detection

Base category: bounding box; Novel category: image label (chaotic names: mixed-supervised, cross-supervised, partially-supervised, weak-shot)

• Judy Hoffman, Sergio Guadarrama, Eric Tzeng, Ronghang Hu, Jeff Donahue, Ross Girshick, Trevor Darrell, Kate Saenko: "LSDA: Large Scale Detection Through Adaptation." NIPS (2014) [paper] [code]
• Joseph Redmon, Ali Farhadi: "YOLO9000: Better, Faster, Stronger." CVPR (2017) [paper] [code]
• Bharat Singh, Hengduo Li, Abhishek Sharma, Larry S. Davis: "R-FCN-3000 at 30fps: Decoupling detection and classification." CVPR (2018) [paper] [code]
• Yan Li, Junge Zhang, Kaiqi Huang, Jianguo Zhang: "Mixed Supervised Object Detection with Robust Objectness Transfer." T-PAMI (2018) [paper] [arXiv]
• Jason Kuen, Federico Perazzi, Zhe Lin, Jianming Zhang, Yap-Peng Tan: "Scaling Object Detection by Transferring Classification Weights." ICCV (2019) [paper] [code]
• Yuanyi Zhong, Jianfeng Wang, Jian Peng, Lei Zhang: "Boosting Weakly Supervised Object Detection with Progressive Knowledge Transfer." ECCV (2020) [paper] [arXiv] [code]
• Ye Guo, Yali Li, Shengjin Wang: "Cs-r-fcn: Cross-supervised Learning for Large-scale Object Detection." ICASSP (2020) [arXiv]
• Zitian Chen, Zhiqiang Shen, Jiahui Yu, Erik Learned-Miller: "Cross-Supervised Object Detection." arXiv preprint arXiv:2006.15056 (2020). [arXiv]
• Yan Liu, Zhijie Zhang, Li Niu, Junjie Chen, Liqing Zhang: "Mixed Supervised Object Detection by Transferring Mask Prior and Semantic Similarity." NeurIPS (2021) [code]

Weak-Shot Semantic Segmentation

Base category: semantic mask; Novel category: image label (weak-shot)

• Siyuan Zhou, Li Niu, Jianlou Si, Chen Qian, Liqing Zhang: "Weak-shot Semantic Segmentation by Transferring Semantic Affinity and Boundary." arXiv preprint arXiv:2110.01519 (2021). [arXiv]

Weak-Shot Instance Segmentation

Base category: instance mask; Novel category: bounding box (partially-supervised)

• Ronghang Hu, Piotr Dollar, Kaiming He, Trevor Darrell, Ross Girshick: "Learning to Segment Every Thing." CVPR (2018) [paper] [code]
• Weicheng Kuo, Anelia Angelova, Jitendra Malik, Tsung-Yi Lin: "ShapeMask: Learning to Segment Novel Objects by Refining Shape Priors." ICCV (2019) [paper] [arXiv]
• Yanzhao Zhou, Xin Wang, Jianbin Jiao, Trevor Darrell, Fisher Yu: "Learning Saliency Propagation for Semi-Supervised Instance Segmentation." CVPR (2020) [paper] [code]
• Qi Fan, Lei Ke, Wenjie Pei, Chi-Keung Tang, Yu-Wing Tai: "Commonality-Parsing Network across Shape and Appearance for Partially Supervised Instance Segmentation." ECCV (2020) [arXiv] [code]
• David Biertimpel, Sindi Shkodrani, Anil S. Baslamisli, Nora Baka: "Prior to Segment: Foreground Cues for Weakly Annotated Classes in Partially Supervised Instance Segmentation." arXiv preprint arXiv:2011.11787 (2020) [arXiv] [code]
• Vighnesh Birodkar, Zhichao Lu, Siyang Li, Vivek Rathod, Jonathan Huang: "The Surprising Impact of Mask-head Architecture on Novel Class Segmentation." arXiv preprint arXiv:2104.00613 (2021) [arXiv] [code]