mlsanityData to 14 Oct 2026

BOCCHI: A More Realistic and Challenging Benchmark for Local Motion Blur Detection with MSDCT-UNet

arXiv 2607.10427 ↗First posted 11 Jul 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Local motion blur detection requires pixel-level localization of blurred regions. Existing benchmarks let models rely on gradient shortcuts that fail to transfer. We introduce BOCCHI (Blurred Objects Captured across Cameras with Human-annotated Imagery), a real-captured benchmark whose sharp regions overlap the blur gradient distribution and defeat these shortcuts, and propose MSDCT-UNet (Multi-Scale Discrete Cosine Transform UNet), a frequency-aware encoder-decoder injecting multi-scale DCT priors through DCT Attention and FiLM. MSDCT-UNet ranks first in in-domain mIoU and boundary localization on BOCCHI, and BOCCHI-trained models outperform every other training source on cross-dataset transfer with only 633 training images. Our project page is available at https://brianchen1120.github.io/project/bocchi/.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 31 of its 53 references that are in the map. First posted on arXiv 11 Jul 2026.

Builds on

31 of 53 references.

References · 31

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  2. 2017Squeeze-and-Excitation Networkscited by 4,789
  3. 2022A ConvNet for the 2020scited by 3,579
  4. 2018Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentationcited by 3,223
  5. 2016Pyramid Scene Parsing Networkcited by 2,479
  6. 2018CBAM: Convolutional Block Attention Modulecited by 2,477
  7. 2017FiLM: Visual Reasoning with a General Conditioning Layercited by 2,333
  8. 2021Restormer: Efficient Transformer for High-Resolution Image Restorationcited by 1,627
  9. 2019Deep High-Resolution Representation Learning for Visual Recognitioncited by 1,420
  10. 2016Deep Multi-scale Convolutional Neural Network for Dynamic Scene Deblurringcited by 727
  11. 2022Simple Baselines for Image Restorationcited by 708
  12. 2020BiSeNet V2: Bilateral Network with Guided Aggregation for Real-time Semantic Segmentationcited by 294
19 more
  1. 2022SegNeXt: Rethinking Convolutional Attention Design for Semantic Segmentationcited by 252
  2. 2020Focal Frequency Loss for Image Reconstruction and Synthesiscited by 206
  3. 2020Learning in the Frequency Domaincited by 198
  4. 2020FcaNet: Frequency Channel Attention Networkscited by 179
  5. 2015Learning a Convolutional Neural Network for Non-uniform Motion Blur Removalcited by 149
  6. 2022Stripformer: Strip Transformer for Fast Image Deblurringcited by 140
  7. 2019Fast-SCNN: Fast Semantic Segmentation Networkcited by 125
  8. 2021Rethinking BiSeNet For Real-time Semantic Segmentationcited by 124
  9. 2016From Motion Blur to Motion Flow: a Deep Learning Solution for Removing Heterogeneous Motion Blurcited by 81
  10. 2023Adaptive Frequency Filters As Efficient Global Token Mixerscited by 45
  11. 2022Real-World Deep Local Motion Deblurringcited by 21
  12. 2017Spatially-Varying Blur Detection Based on Multiscale Fused and Sorted Transform Coefficients of Gradient Magnitudescited by 15
  13. 2020Defocus Blur Detection via Depth Distillationcited by 12
  14. 2024Learning to Control Camera Exposure via Reinforcement Learningcited by 9
  15. 2024AlignSAM: Aligning Segment Anything Model to Open Context via Reinforcement Learningcited by 8
  16. 2023Adaptive Window Pruning for Efficient Local Motion Deblurringcited by 6
  17. 2020Dense Dual-Path Network for Real-time Semantic Segmentationcited by 3
  18. 2026RL-AWB: Deep Reinforcement Learning for Auto White Balance Correction in Low-Light Night-time Scenescited by 2
  19. 2020MSDU-net: A Multi-Scale Dilated U-net for Blur Detectioncited by 1

Built on by

Not cited yet.