Learning Clustering-Based Linear Mappings for Quantization Noise Removal

Abstract : This paper describes a novel scheme to reduce the quantization noise of compressed videos and improve the overall coding performances. The proposed scheme first consists in clustering noisy patches of the compressed sequence. Then, at the encoder side, linear mappings are learned for each cluster between the noisy patches and the corresponding source patches. The linear mappings are then transmitted to the decoder where they can be applied to perform de-noising. The method has been tested with the HEVC standard, leading to a bitrate saving of up to 9.63%.
Document type :
Conference papers
Complete list of metadatas

Cited literature [33 references]  Display  Hide  Download

https://hal.inria.fr/hal-01317625
Contributor : Martin Alain <>
Submitted on : Wednesday, May 18, 2016 - 4:22:44 PM
Last modification on : Thursday, February 7, 2019 - 4:33:32 PM
Long-term archiving on : Friday, August 19, 2016 - 10:47:20 AM

File

Clustering_based_denoising.pdf
Files produced by the author(s)

Identifiers

Citation

Martin Alain, Christine Guillemot, Dominique Thoreau, Philippe Guillotel. Learning Clustering-Based Linear Mappings for Quantization Noise Removal. IEEE International Conference on Image Processing (ICIP) 2016, Sep 2016, Phoenix, United States. ⟨10.1109/ICIP.2016.7533151⟩. ⟨hal-01317625⟩

Share

Metrics

Record views

944

Files downloads

280