Abstract:JPEG is one of the most widely used lossy image compression algorithms, but artifacts are generated during compression. Various artifact reduction methods have been proposed, and many of them, especially deep learning-based approaches, showed promising performance. However, one major drawback that limited their deployment and application is their cumbersome and complicated model. To remedy this problem, we propose a simple and efficient network named Efficient Artifact Reduction Network. To achieve efficiency, we consider enlarging the receptive field and preserving pixel-wise information as significant concerns. On the one hand, we notice choosing a proper down-sampling ratio is important, as the down-sampling operation is a trade-off between these two aspects. On the other hand, we design a Large Kernel Depthwise Separable Convolution block that considers both aspects. For flexibility over different compression qualities, which is the focus of research in recent years, we design a Half Adaptive Instance Normalization-based approach that elegantly integrates information of the Quantization Matrix into the feature map. It adaptively normalizes half of the channels in the Encoder to embed the compression quality information and precise pixel-wise information is preserved through the other half channels. We also design a scalable architecture inspired by prior works to enable a post-training balance between computational cost and restoration performance. Experiments on various datasets show that our network achieves state-of-the-art restoration performance with much fewer parameters and less computational cost.

A Low-Complexity Algorithm for JPEG-LS-Based RAW Domain Compression

UH-JLS: A Parallel Ultra-High Throughput JPEG-LS Encoding Architecture for Lossless Image Compression

A Low Power, Fully Pipelined JPEG-LS Encoder for Lossless Image Compression

Lossless Recompression of JPEG Images Using Transform Domain Intra Prediction

An Improved Algorithm of JPEG-LS for Near-lossless Compression of Remote Sensing Images

Efficient Learned Lossless JPEG Recompression

Lossless compression of JPEG coded photo albums

Context-Based Lossless Compression of Mosaic Image with Bayer Pattern

Low-complexity Transform Network Architecture for JPEG AI Image Codec.

The LOCO-I lossless image compression algorithm: principles and standardization into JPEG-LS

Adaptive Pipeline Hardware Architecture Design and Implementation for Image Lossless Compression/Decompression Based on JPEG-LS

Unified learning-based lossy and lossless JPEG recompression

EARN: Toward Efficient and Robust JPEG Compression Artifact Reduction

Fuzzy Neural Network Based Prediction Coding for Bayer Pattern Image

Learned Lossless Compression for JPEG via Frequency-Domain Prediction

High efficiency lossless image recompression algorithm with asymmetric numeral systems for real-time mobile application

Design and Implementation of a Low-Power JPEG Compressor IC

A Low-Complexity and High-Quality Image Compression Method for Digital Cameras

A LowPower, Fully Pipelined JPEG-LSEncoder Forlossless ImageCompression

Lossless Compression of JPEG Coded Photo Collections.

An Adaptive Predictive Coding Based on Image Segmentation for Lossless Compression of Ultrasonic Well Logging Images