Learned JPEG Compression for DNN Vision

• URL: http://arxiv.org/abs/2606.16185v1
• Date: Mon, 15 Jun 2026 03:56:45 GMT
• Title: Learned JPEG Compression for DNN Vision
• Abstract summary: In the era of artificial intelligence (AI), a substantial portion of image data, often compressed by JPEG, is and will continue to be consumed by deep neural networks (DNNs) instead of humans.<n>We propose learned JPEG compression for DNN vision (J4D), a novel training framework for determining JPEG encoding parameters to minimize compression rate while maximizing inference performance.
• Score: 15.86593426643386
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: JPEG, a lossy image compression technique designed for human viewers, has maintained its dominance for decades. However, in the era of artificial intelligence (AI), a substantial portion of image data, often compressed by JPEG, is and will continue to be consumed by deep neural networks (DNNs) instead of humans, thus creating a need to optimize JPEG for DNN inference performance. To this end, we propose learned JPEG compression for DNN vision (J4D), a novel training framework for determining JPEG encoding parameters to minimize compression rate while maximizing DNN inference performance. The major challenge of solving this optimization problem lies in representing the JPEG codec and compression rate in closed form. By incorporating a differentiable soft quantizer based on a probabilistic quantization scheme, we not only obtain a differentiable proxy for the JPEG codec, but are also able to compute the entropy of the coded source analytically, which is a close estimate of the actual compression rate. Equipped with both the differentiable JPEG codec and the information-theoretic rate estimator, we are then able to solve the aforementioned optimization problem with backpropagation. After training, the learned encoding parameters will be subsequently used in actual JPEG encoding based on probabilistic quantization. Extensive experimental results across multiple datasets and DNN architectures demonstrate that J4D consistently and significantly outperforms the default JPEG and other competitive JPEG codecs optimized for DNNs. Notably, compared to the default JPEG, J4D achieves an increase in accuracy by as much as 11.60% at the same rate, or a reduction of compression rate up to 80.05% at the same accuracy. Additionally, with the help of J4D, we show the potential to design universal JPEG encoding parameters for various DNN architectures for the first time.

Related papers

• JPEG Inspired Deep Learning [4.958744940097937]
  Well-crafted JPEG compression can actually improve the performance of deep learning (DL)<n>We propose JPEG-DL, a novel DL framework that prepends any underlying DNN architecture with a trainable JPEG compression layer.<n>Our code is available on https://github.com/AhmedHussKhalifa/Inspired-DL.git.
  arXiv  Detail & Related papers  (2024-10-09T17:23:54Z)
• JDEC: JPEG Decoding via Enhanced Continuous Cosine Coefficients [17.437568540883106]
  We propose a practical approach to JPEG image decoding, utilizing a local implicit neural representation with continuous cosine formulation. Our proposed network achieves state-of-the-art performance in flexible color image JPEG artifact removal tasks.
  arXiv  Detail & Related papers  (2024-04-03T03:28:04Z)
• Unified learning-based lossy and lossless JPEG recompression [15.922937139019547]
  We propose a unified lossly and lossless JPEG recompression framework, which consists of learned quantization table and Markovian hierarchical variational autoencoders. Experiments show that our method can achieve arbitrarily low distortion when the JPEG is close to the upper bound.
  arXiv  Detail & Related papers  (2023-12-05T12:07:27Z)
• Learned Lossless Compression for JPEG via Frequency-Domain Prediction [50.20577108662153]
  We propose a novel framework for learned lossless compression of JPEG images. To enable learning in the frequency domain, DCT coefficients are partitioned into groups to utilize implicit local redundancy. An autoencoder-like architecture is designed based on the weight-shared blocks to realize entropy modeling of grouped DCT coefficients.
  arXiv  Detail & Related papers  (2023-03-05T13:15:28Z)
• Neural JPEG: End-to-End Image Compression Leveraging a Standard JPEG Encoder-Decoder [73.48927855855219]
  We propose a system that learns to improve the encoding performance by enhancing its internal neural representations on both the encoder and decoder ends. Experiments demonstrate that our approach successfully improves the rate-distortion performance over JPEG across various quality metrics.
  arXiv  Detail & Related papers  (2022-01-27T20:20:03Z)
• Towards Flexible Blind JPEG Artifacts Removal [73.46374658847675]
  We propose a flexible blind convolutional neural network, namely FBCNN, that can predict the adjustable quality factor to control the trade-off between artifacts removal and details preservation. Our proposed FBCNN achieves favorable performance against state-of-the-art methods in terms of both quantitative metrics and visual quality.
  arXiv  Detail & Related papers  (2021-09-29T17:12:10Z)
• Learning to Improve Image Compression without Changing the Standard Decoder [100.32492297717056]
  We propose learning to improve the encoding performance with the standard decoder. Specifically, a frequency-domain pre-editing method is proposed to optimize the distribution of DCT coefficients. We do not modify the JPEG decoder and therefore our approach is applicable when viewing images with the widely used standard JPEG decoder.
  arXiv  Detail & Related papers  (2020-09-27T19:24:42Z)
• Quantization Guided JPEG Artifact Correction [69.04777875711646]
  We develop a novel architecture for artifact correction using the JPEG files quantization matrix. This allows our single model to achieve state-of-the-art performance over models trained for specific quality settings.
  arXiv  Detail & Related papers  (2020-04-17T00:10:08Z)
• Learning Better Lossless Compression Using Lossy Compression [100.50156325096611]
  We leverage the powerful lossy image compression algorithm BPG to build a lossless image compression system. We model the distribution of the residual with a convolutional neural network-based probabilistic model that is conditioned on the BPG reconstruction. Finally, the image is stored using the concatenation of the bitstreams produced by BPG and the learned residual coder.
  arXiv  Detail & Related papers  (2020-03-23T11:21:52Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.