English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Learned image compression research has achieved state-of-the-art compressionperformance with auto-encoder based neural network architectures, where theimage is mapped via convolutional neural networks (CNN) into a latentrepresentation that is quantized and processed again with CNN to obtain thereconstructed image. CNN operate on entire input images. On the other hand,traditional state-of-the-art image and video compression methods process imageswith a block-by-block processing approach for various reasons. Very recently,work on learned image compression with block based approaches have alsoappeared, which use the auto-encoder architecture on large blocks of the inputimage and introduce additional neural networks that perform intra/spatialprediction and deblocking/post-processing functions. This paper explores analternative learned block-based image compression approach in which neither anexplicit intra prediction neural network nor an explicit deblocking neuralnetwork is used. A single auto-encoder neural network with block-level maskedconvolutions is used and the block size is much smaller (8x8). By usingblock-level masked convolutions, each block is processed using reconstructedneighboring left and upper blocks both at the encoder and decoder. Hence, themutual information between adjacent blocks is exploited during compression andeach block is reconstructed using neighboring blocks, resolving the need forexplicit intra prediction and deblocking neural networks. Since the exploredsystem is a closed loop system, a special optimization procedure, theasymptotic closed loop design, is used with standard stochastic gradientdescent based training. The experimental results indicate competitive imagecompression performance.

End-to-End Learned Block-Based Image Compression with Block-Level Masked Convolutions and Asymptotic Closed Loop Training