Bridging the Gap Between Image Coding for Machines and Humans

Image coding for machines (ICM) aims at reducing the bitrate required to represent an image while minimizing the drop in machine vision analysis accuracy. In many use cases, such as surveillance, it is also important that the visual quality is not drastically deteriorated by the compression process....

Full description

Saved in:
Bibliographic Details
Published in:Proceedings - International Conference on Image Processing pp. 3411 - 3415
Main Authors: Le, Nam, Zhang, Honglei, Cricri, Francesco, Youvalari, Ramin G., Rezazadegan Tavakoli, Hamed, Aksu, Emre, Hannuksela, Miska M., Rahtu, Esa
Format: Conference Proceeding
Language:English
Published: IEEE 16.10.2022
Subjects:
Codecs
finetuning
GANs
Image coding
Image coding for machines
Machine vision
Measurement
Surveillance
Training
VCM
Visualization
ISSN:2381-8549
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Image coding for machines (ICM) aims at reducing the bitrate required to represent an image while minimizing the drop in machine vision analysis accuracy. In many use cases, such as surveillance, it is also important that the visual quality is not drastically deteriorated by the compression process. Recent works on using neural network (NN) based ICM codecs have shown significant coding gains against traditional methods; however, the decompressed images, especially at low bitrates, often contain checkerboard artifacts. We propose an effective decoder finetuning scheme based on adversarial training to significantly enhance the visual quality of ICM codecs, while preserving the machine analysis accuracy, without adding extra bitcost or parameters at the inference phase. The results show complete removal of the checkerboard artifacts at the negligible cost of −1.6% relative change in task performance score. In the cases where some amount of artifacts is tolerable, such as when machine consumption is the primary target, this technique can enhance both pixel-fidelity and feature-fidelity scores without losing task performance.
ISSN:2381-8549
DOI:10.1109/ICIP46576.2022.9897916