Learned Lossless Image Compression with A Hyperprior and Discretized Gaussian Mixture Likelihoods

Cheng, Zhengxue; Sun, Heming; Takeuchi, Masaru; Katto, Jiro

doi:10.1109/icassp40776.2020.9053413

Cited by 9 publications

(3 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…al [34] and Cheng et. al [17] suggested that compressing residual of compressed image with traditional methods is also feasible for lossless image compression with end-to-end models.…”

Section: Learned Lossless Compressionmentioning

confidence: 99%

Learned Lossless Compression for JPEG via Frequency-Domain Prediction

Luo¹,

Li²,

Dai³

et al. 2023

Preprint

View full text Add to dashboard Cite

JPEG images can be further compressed to enhance the storage and transmission of large-scale image datasets. Existing learned lossless compressors for RGB images cannot be well transferred to JPEG images due to the distinguishing distribution of DCT coefficients and raw pixels. In this paper, we propose a novel framework for learned lossless compression of JPEG images that achieves endto-end optimized prediction of the distribution of decoded DCT coefficients. 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 and independently compress the priors. We attempt to realize learned lossless compression of JPEG images in the frequency domain. Experimental results demonstrate that the proposed framework achieves superior or comparable performance in comparison to most recent lossless compressors with handcrafted context modeling for JPEG images.

show abstract

“…al [34] and Cheng et. al [17] suggested that compressing residual of compressed image with traditional methods is also feasible for lossless image compression with end-to-end models.…”

Section: Learned Lossless Compressionmentioning

confidence: 99%

Learned Lossless Compression for JPEG via Frequency-Domain Prediction

Luo¹,

Li²,

Dai³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…When it comes to lossy picture compression, HyperPrior is one of the best methods available. The training procedure is sped up by adding an L2 norm component to the loss function, as proposed in paper [22], which generalizes the supernormal prior from the lossy model to lossless compression. This work also discusses the use of Gaussian mixture probabilities to construct adaptable and flexible texture models and evaluates various parametric models for hidden codes.…”

Section: Literature Reviewmentioning

confidence: 99%

Medical image compression and restoration using a hybrid neural network combining Huffman and meta-heuristic techniques

Khalifeh,

Taghizadeh,

Ghanbarian

et al. 2023

Preprint

View full text Add to dashboard Cite

In order to compress medical pictures for long-term storage, two methods are used in this work. The first step is to use a neural network–based categorization system to simplify images using a hierarchical modeling technique. The Huffman cipher is then used to compress the reduced images. In the second method, a deep neural network is trained to make predictions. This method can potentially reduce the amount of data needed to describe a picture by using a trained neural network to make intelligent guesses about the location of individual pixels. Huffman compression is used to encrypt the remaining data. By using an improved spatial filtering method to the picture data, we can decode it and then use meta-heuristic algorithms like gray wolf optimization (GWO) and wild horse optimization (WHO) to rebuild the image. Without sacrificing data compression efficacy, this paves the way for a more practical implementation of the proposed techniques in cases when outcomes are uncertain. Images can be simplified using the suggested approaches, leading to faster decoding. Afterwards, performance metrics were taken and evaluated following predetermined daily procedures. The suggested approaches outperformed state-of-the-art deep learning-based systems in compressing medical images while maintaining an exceptionally high quality level.

show abstract

“…By defining the number of nodes at the end of an encoding subnet, the image is transformed into a feature vector with a small number of elements, thus reducing the redundancy of the image. So far, various kinds of deep learning networks, including recurrent neural networks (RNNs) [5,6], convolutional neural networks (CNNs) [7,8,9,10,11,12,13], and generative adversarial networks (GANs) [14,15], have been explored for image compression. Although these methods have achieved great results in certain datasets, there are still some shortcomings.…”

Section: Introductionmentioning

confidence: 99%

Sparse Flow Adversarial Model For Robust Image Compression

Zhao

Yang

Zhi

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Existing learned-based image compression methods have shown impressive performance. However, most of them rely on the consistency of distribution between training images and test images, which limits the robustness of the trained model. In this paper, we propose a novel compression method called sparse flow adversarial model (SFAM). SFAM employs a deep generative framework to learn a reversible and stable mapping between image distributions, thus it can work in varied scenes for robust compression. Moreover, a sparse adversarial map is introduced into SFAM, to constrain the SFAM to generate more sparser features for efficient compression. Extensive experiments are conducted on different datasets, in which the effectiveness and robustness of the proposed method is verified. Meanwhile, SFAM is trained only once and it can work well on three different datasets, which also proves the robustness of the proposed SFAM.

show abstract

Learned Lossless Image Compression with A Hyperprior and Discretized Gaussian Mixture Likelihoods

Cited by 9 publications

References 23 publications

Learned Lossless Compression for JPEG via Frequency-Domain Prediction

Learned Lossless Compression for JPEG via Frequency-Domain Prediction

Medical image compression and restoration using a hybrid neural network combining Huffman and meta-heuristic techniques

Sparse Flow Adversarial Model For Robust Image Compression

Contact Info

Product

Resources

About