mlsanityData to 7 Oct 2026

ECSIC: Epipolar Cross Attention for Stereo Image Compression

arXiv 2307.10284 ↗First posted 18 Jul 2023eess.IV cs.CV cs.LGcited by 8

Not on the map. Why?

Abstract (arXiv)

In this paper, we present ECSIC, a novel learned method for stereo image compression. Our proposed method compresses the left and right images in a joint manner by exploiting the mutual information between the images of the stereo image pair using a novel stereo cross attention (SCA) module and two stereo context modules. The SCA module performs cross-attention restricted to the corresponding epipolar lines of the two images and processes them in parallel. The stereo context modules improve the entropy estimation of the second encoded image by using the first image as a context. We conduct an extensive ablation study demonstrating the effectiveness of the proposed modules and a comprehensive quantitative and qualitative comparison with existing methods. ECSIC achieves state-of-the-art performance in stereo image compression on the two popular stereo image datasets Cityscapes and InStereo2k while allowing for fast encoding and decoding.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 32 of its 50 references that are in the map. First posted on arXiv 18 Jul 2023.

Builds on

32 of 50 references.

References · 32

  1. 2017Attention Is All You Needcited by 59,932
  2. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,427
  3. 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,590
  4. 2018Self-Attention with Relative Position Representationscited by 1,112
  5. 2018Variational image compression with a scale hyperpriorcited by 910
  6. 2016End-to-end Optimized Image Compressioncited by 673
  7. 2018Joint Autoregressive and Hierarchical Priors for Learned Image Compressioncited by 642
  8. 2020Learned Image Compression with Discretized Gaussian Mixture Likelihoods and Attention Modulescited by 570
  9. 2020High-Fidelity Generative Image Compressioncited by 331
  10. 2022ELIC: Efficient Learned Image Compression with Unevenly Grouped Space-Channel Contextual Adaptive Codingcited by 296
  11. 2017Lossy Image Compression with Compressive Autoencoderscited by 289
  12. 2020Channel-wise Autoregressive Entropy Models for Learned Image Compressioncited by 255
20 more
  1. 2018Generative Adversarial Networks for Extreme Learned Image Compressioncited by 213
  2. 2021Checkerboard Context Model for Efficient Learned Image Compressioncited by 193
  3. 2015Variable Rate Image Compression with Recurrent Neural Networkscited by 168
  4. 2022The Devil Is in the Details: Window-based Attention for Image Compressioncited by 141
  5. 2018Conditional Probability Models for Deep Image Compressioncited by 128
  6. 2021Enhanced Invertible Encoding for Learned Image Compressioncited by 102
  7. 2022Entroformer: A Transformer-based Entropy Model for Learned Image Compressioncited by 102
  8. 2022Lossy Compression with Gaussian Diffusioncited by 87
  9. 2020Improving Inference for Neural Image Compressioncited by 81
  10. 2020Causal Contextual Prediction for Learned Image Compressioncited by 70
  11. 2018Deep Generative Models for Distribution-Preserving Lossy Compressioncited by 59
  12. 2021Joint Global and Local Hierarchical Priors for Learned Image Compressioncited by 52
  13. 2021Soft then Hard: Rethinking the Quantization in Neural Image Compressioncited by 50
  14. 2020Learning Accurate Entropy Model with Global Reference for Image Compressioncited by 42
  15. 2022Contextformer: A Transformer with Spatio-Channel Attention for Context Modeling in Learned Image Compressioncited by 42
  16. 2022PO-ELIC: Perception-Oriented Efficient Learned Image Codingcited by 35
  17. 2019DSIC: Deep Stereo Image Compressioncited by 20
  18. 2022Neural Distributed Image Compression with Cross-Attention Feature Alignmentcited by 15
  19. 2023LDMIC: Learning-based Distributed Multi-view Image Codingcited by 13
  20. 2021L3C-Stereo: Lossless Compression for Stereo Imagescited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 3; 2025: 3; 2026: 2.

8 papers.

  1. 2024Bidirectional Stereo Image Compression with Cross-Dimensional Entropy Modelcited by 8
  2. 2024FCNR: Fast Compressive Neural Representation of Visualization Imagescited by 6
  3. 2024CAMSIC: Content-aware Masked Image Modeling Transformer for Stereo Image Compressioncited by 3
  4. 2025Neural Stereo Video Compression with Hybrid Disparity Compensationcited by 2
  5. 2025Synthesizing Consistent Novel Views via 3D Epipolar Attention without Re-Trainingcited by 1
  6. 2025Accelerating Learned Image Compression Through Modeling Neural Training Dynamicscited by 1
  7. 2026Distributed Image Compression with Multimodal Side Information at Extremely Low Bitratescited by 0
  8. 2026Echo4DIR: 4D Implicit Heart Reconstruction from 2D Echocardiography Videoscited by 0