Learned Image Coding for Machines: A Content-Adaptive Approach

  title={Learned Image Coding for Machines: A Content-Adaptive Approach},
  author={Nam Do-Hoang Le and Honglei Zhang and Francesco Cricri and Ramin Ghaznavi-Youvalari and Hamed Rezazadegan Tavakoli and Esa Rahtu},
Today, according to the Cisco Annual Internet Report (20182023), the fastest-growing category of Internet traffic is machine-to-machine communication. In particular, machineto-machine communication of images and videos represents a new challenge and opens up new perspectives in the context of data compression. One possible solution approach consists of adapting current human-targeted image and video coding standards to the use case of machine consumption. Another approach consists of developing… 
1 Citations

Figures and Tables from this paper

Video Coding for Machine: Compact Visual Representation Compression for Intelligent Collaborative Analytics
A novel visual information compression for the analytics taxonomy problem is investigated to strengthen the capability of compact visual representations extracted from multiple tasks for visual analytics and a new perspective of task relationships versus compression is revisited.


Image Coding For Machines: an End-To-End Learned Approach
This paper proposes an image codec for machines which is neural network (NN) based and end-to-end learned, and shows that its NN-based codec outperforms the state-of-the-art Versa-tile Video Coding standard on the object detection and instance segmentation tasks.
Semantic-Preserving Image Compression
This work applies an approach to compress visual data in order to maximize performance on a given analytic task on an image classification problem and shows that for a given level of compression, it achieves higher classification accuracy than that obtained by performing classification on images compressed using JPEG.
Video Coding for Machines with Feature-Based Rate-Distortion Optimization
This paper replaces the pixel-based distortion metrics in conventional RDO of VTM-8.0 with distortion metrics calculated in the feature space created by the first layers of a neural network, which results in coding gains for the proposed HFRDO of up 9.95% compared to conventional VTM.
Content Adaptive Optimization for Neural Image Compression
This work introduces an iterative procedure which adapts the latent representation to the specific content the authors wish to compress while keeping the parameters of the network and the predictive model fixed and shows that this allows for an overall increase in rate-distortion performance, independently of the specific architecture used.
Compressing Weight-updates for Image Artifacts Removal Neural Networks
This paper presents a novel approach for fine-tuning a decoder-side neural network in the context of image compression, such that the weight-updates are better compressible and achieves reconstruction quality which is on-par or slightly superior to a traditional codec, at comparable bitrates.
Lossless Image Compression Using a Multi-scale Progressive Statistical Model
This paper has developed a flexible mechanism where the processing order of the pixels can be adjusted easily and outperforms the state-of-the-art lossless image compression methods on two large benchmark datasets by a significant margin.
L2C – Learning to Learn to Compress
This paper proposes a new training paradigm for learned image compression, which is based on meta-learning, and proposes to overfit and cluster the bias terms of the decoder on training image patches, so that at inference time the optimal content-specific bias terms can be selected at encoder-side.
DeepSIC: Deep Semantic Image Compression
This paper proposes a concept called Deep Semantic Image Compression (DeepSIC) and puts forward two novel architectures that aim to reconstruct the compressed image and generate corresponding semantic representations at the same time.
Very deep convolutional neural network based image classification using small training sample size
The results show that the very deep CNN can be used to fit small datasets with simple and proper modifications and don't need to re-design specific small networks.
Mask R-CNN
This work presents a conceptually simple, flexible, and general framework for object instance segmentation that outperforms all existing, single-model entries on every task, including the COCO 2016 challenge winners.