mlsanityData to 7 Oct 2026

When W4A4 Breaks Camouflaged Object Detection: Token-Group Dual-Constraint Activation Quantization

arXiv 2604.16855 ↗First posted 18 Apr 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Camouflaged object detection (COD) segments objects that intentionally blend with the background, so predictions depend on subtle texture and boundary cues. COD is often needed under tight on-device memory and latency budgets, making low-bit inference highly desirable. However, COD is unusually hard to quantize aggressively. We study post-training W4A4 quantization of Transformer-based COD and find a task-specific cliff: heavy-tailed background tokens dominate a shared activation range, inflating the step size and pushing weak-but-structured boundary cues into the zero bin. This exposes a token-local bottleneck -- remove cross-token range domination and bound the zero-bin mass under 4-bit activations. To address this, we introduce COD-TDQ, a COD-aware Token-group Dual-constraint activation Quantization method. COD-TDQ addresses this token-local bottleneck with two coupled steps: Direct-Sum Token-Group (DSTG) assigns token-group scales to suppress cross-token range domination, and Dual-Constraint Range Projection (DCRP) projects each token-group clip range to keep the step-to-dispersion ratio and the zero-bin mass bounded. Across four COD benchmarks and two baseline models (CFRN and ESCNet), COD-TDQ consistently achieves an Sα score more than 0.12 higher than that of the state-of-the-art quantization method without retraining. The code is available at https://github.com/MCG-NKU/nku-model-compre.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 32 of its 59 references that are in the map. First posted on arXiv 18 Apr 2026.

Builds on

32 of 59 references.

References · 32

  1. 2017Attention Is All You Needcited by 59,932
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  3. 2022GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformerscited by 1,838
  4. 2022SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Modelscited by 1,338
  5. 2020Up or Down? Adaptive Rounding for Post-Training Quantizationcited by 600
  6. 2021Anabranch Network for Camouflaged Object Segmentationcited by 227
  7. 2023Parameter-Efficient Fine-Tuning Methods for Pretrained Language Models: A Critical Review and Assessmentcited by 227
  8. 2022Visual Attention Networkcited by 210
  9. 2021Simultaneously Localize, Segment and Rank the Camouflaged Objectscited by 198
  10. 2022QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantizationcited by 175
  11. 2021PTQ4ViT: Post-training quantization for vision transformers with twin uniform quantizationcited by 173
  12. 2022RepQ-ViT: Scale Reparameterization for Post-Training Quantization of Vision Transformerscited by 134
20 more
  1. 2023Segment Anything Is Not Always Perfect: An Investigation of SAM on Different Real-world Applicationscited by 121
  2. 2025Pretraining Large Language Models with NVFP4cited by 67
  3. 2022NoisyQuant: Noisy Bias-Enhanced Post-Training Activation Quantization for Vision Transformerscited by 59
  4. 2022CamoFormer: Masked Separable Attention for Camouflaged Object Detectioncited by 58
  5. 2024PTQ4SAM: Post-Training Quantization for Segment Anythingcited by 32
  6. 2024A Simple yet Effective Network based on Vision Transformer for Camouflaged Object and Salient Object Detectioncited by 23
  7. 2024Instance-Aware Group Quantization for Vision Transformerscited by 23
  8. 2025FIMA-Q: Post-Training Quantization for Vision Transformers by Fisher Information Matrix Approximationcited by 16
  9. 2023Open-Vocabulary Camouflaged Object Segmentationcited by 14
  10. 2026ARCQuant: Boosting NVFP4 Quantization with Augmented Residual Channels for LLMscited by 14
  11. 2025Improving SAM for Camouflaged Object Detection via Dual Stream Adapterscited by 13
  12. 2024When SAM2 Meets Video Camouflaged Object Segmentation: A Comprehensive Evaluation and Adaptationcited by 11
  13. 2025Controllable-LPMoE: Adapting to Challenging Object Segmentation via Dynamic Local Priors from Mixture-of-Expertscited by 9
  14. 2024ADFQ-ViT: Activation-Distribution-Friendly Post-Training Quantization for Vision Transformerscited by 7
  15. 2024Rethinking Detecting Salient and Camouflaged Objects in Unconstrained Scenescited by 7
  16. 2025Evaluating SAM2 for Video Semantic Segmentationcited by 4
  17. 2024Towards Real Zero-Shot Camouflaged Object Segmentation without Camouflaged Annotationscited by 3
  18. 2025Open-Vocabulary Camouflaged Object Segmentation with Cascaded Vision Language Modelscited by 3
  19. 2025GAPNet: A Lightweight Framework for Image and Video Salient Object Detection via Granularity-Aware Paradigmcited by 1
  20. 2025Post-Training Quantization via Residual Truncation and Zero Suppression for Diffusion Modelscited by 1

Built on by

Not cited yet.