mlsanityData to 7 Oct 2026

Boosting Few-Shot Segmentation via Instance-Aware Data Augmentation and Local Consensus Guided Cross Attention

arXiv 2401.09866 ↗First posted 18 Jan 2024cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Few-shot segmentation aims to train a segmentation model that can fast adapt to a novel task for which only a few annotated images are provided. Most recent models have adopted a prototype-based paradigm for few-shot inference. These approaches may have limited generalization capacity beyond the standard 1- or 5-shot settings. In this paper, we closely examine and reevaluate the fine-tuning based learning scheme that fine-tunes the classification layer of a deep segmentation network pre-trained on diverse base classes. To improve the generalizability of the classification layer optimized with sparsely annotated samples, we introduce an instance-aware data augmentation (IDA) strategy that augments the support images based on the relative sizes of the target objects. The proposed IDA effectively increases the support set's diversity and promotes the distribution consistency between support and query images. On the other hand, the large visual difference between query and support images may hinder knowledge transfer and cripple the segmentation performance. To cope with this challenge, we introduce the local consensus guided cross attention (LCCA) to align the query feature with support features based on their dense correlation, further improving the model's generalizability to the query image. The significant performance improvements on the standard few-shot segmentation benchmarks PASCAL-5i and COCO-20i verify the efficacy of our proposed method.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 42 of its 53 references that are in the map. First posted on arXiv 18 Jan 2024.

Builds on

42 of 53 references.

References · 42

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  4. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
  5. 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,845
  6. 2016DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFscited by 3,239
  7. 2017Prototypical Networks for Few-shot Learningcited by 3,001
  8. 2016Pyramid Scene Parsing Networkcited by 2,472
  9. 2016V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentationcited by 2,284
  10. 2016Matching Networks for One Shot Learningcited by 2,246
  11. 2017Rethinking Atrous Convolution for Semantic Image Segmentationcited by 2,154
  12. 2015Multi-Scale Context Aggregation by Dilated Convolutionscited by 1,128
30 more
  1. 2019A Closer Look at Few-shot Classificationcited by 679
  2. 2019PANet: Few-Shot Image Semantic Segmentation with Prototype Alignmentcited by 449
  3. 2017One-Shot Learning for Semantic Segmentationcited by 337
  4. 2019A Baseline for Few-Shot Image Classificationcited by 298
  5. 2018Few-Shot Learning via Embedding Adaptation with Set-to-Set Functionscited by 293
  6. 2020Prior Guided Feature Enrichment Network for Few-Shot Segmentationcited by 271
  7. 2017HPatches: A benchmark and evaluation of handcrafted and learned local descriptorscited by 261
  8. 2019CANet: Class-Agnostic Segmentation Networks with Iterative Refinement and Attentive Few-Shot Learningcited by 217
  9. 2021Adaptive Prototype Learning and Allocation for Few-Shot Segmentationcited by 207
  10. 2021Hypercorrelation Squeeze for Few-Shot Segmentationcited by 193
  11. 2020Prototype Mixture Models for Few-shot Semantic Segmentationcited by 190
  12. 2019Feature Weighting and Boosting for Few-Shot Segmentationcited by 189
  13. 2018Neighbourhood Consensus Networkscited by 173
  14. 2020Part-aware Prototype Network for Few-shot Semantic Segmentationcited by 162
  15. 2018Task-Agnostic Meta-Learning for Few-shot Learningcited by 151
  16. 2020Few-Shot Segmentation Without Meta-Learning: A Good Transductive Inference Is All You Need?cited by 135
  17. 2019Revisiting Local Descriptor based Image-to-Class Measure for Few-shot Learningcited by 121
  18. 2021Simpler is Better: Few-shot Semantic Segmentation with Classifier Weight Transformercited by 116
  19. 2020Feature Pyramid Transformercited by 94
  20. 2015Proposal Flowcited by 57
  21. 2019Hyperpixel Flow: Semantic Correspondence with Multi-layer Neural Featurescited by 56
  22. 2020Learning to Compose Hypercolumns for Visual Correspondencecited by 53
  23. 2020On Episodes, Prototypical Networks, and Few-shot Learningcited by 46
  24. 2021Convolutional Hough Matching Networkscited by 43
  25. 2020Correspondence Networks with Adaptive Neighbourhood Consensuscited by 42
  26. 2021Multi-scale Matching Networks for Semantic Correspondencecited by 32
  27. 2022TransforMatcher: Match-to-Match Attention for Semantic Correspondencecited by 29
  28. 2022Beyond the Prototype: Divide-and-conquer Proxies for Few-shot Segmentationcited by 28
  29. 2021MFNet: Multi-class Few-shot Segmentation Network with Pixel-wise Metric Learningcited by 10
  30. 2023Quaternion-valued Correlation Learning for Few-Shot Semantic Segmentationcited by 10

Built on by

Not cited yet.