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

Recently, numerous end-to-end optimized image compression neural networkshave been developed and proved themselves as leaders in rate-distortionperformance. The main strength of these learnt compression methods is inpowerful nonlinear analysis and synthesis transforms that can be facilitated bydeep neural networks. However, out of operational expediency, most of theseend-to-end methods adopt uniform scalar quantizers rather than vectorquantizers, which are information-theoretically optimal. In this paper, wepresent a novel Lattice Vector Quantization scheme coupled with a spatiallyAdaptive Companding (LVQAC) mapping. LVQ can better exploit the inter-featuredependencies than scalar uniform quantization while being computationallyalmost as simple as the latter. Moreover, to improve the adaptability of LVQ tosource statistics, we couple a spatially adaptive companding (AC) mapping withLVQ. The resulting LVQAC design can be easily embedded into any end-to-endoptimized image compression system. Extensive experiments demonstrate that forany end-to-end CNN image compression models, replacing uniform quantizer byLVQAC achieves better rate-distortion performance without significantlyincreasing the model complexity.

LVQAC: Lattice Vector Quantization Coupled with Spatially Adaptive Companding for Efficient Learned Image Compression