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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Lossy image compression strategies allow for more efficient storage andtransmission of data by encoding data to a reduced form. This is essentialenable training with larger datasets on less storage-equipped environments.However, such compression can cause severe decline in performance of deepConvolution Neural Network (CNN) architectures even when mild compression isapplied and the resulting compressed imagery is visually identical. In thiswork, we apply the lossy JPEG compression method with six discrete levels ofincreasing compression {95, 75, 50, 15, 10, 5} to infrared band (thermal)imagery. Our study quantitatively evaluates the affect that increasing levelsof lossy compression has upon the performance of characteristically diverseobject detection architectures (Cascade-RCNN, FSAF and Deformable DETR) withrespect to varying sizes of objects present in the dataset. When training andevaluating on uncompressed data as a baseline, we achieve maximal mean AveragePrecision (mAP) of 0.823 with Cascade R-CNN across the FLIR dataset,outperforming prior work. The impact of the lossy compression is more extremeat higher compression levels (15, 10, 5) across all three CNN architectures.However, re-training models on lossy compressed imagery notably amelioratedperformances for all three CNN models with an average increment of ~76% (athigher compression level 5). Additionally, we demonstrate the relativesensitivity of differing object areas {tiny, small, medium, large} with respectto the compression level. We show that tiny and small objects are moresensitive to compression than medium and large objects. Overall, Cascade R-CNNattains the maximal mAP across most of the object area categories.

Lost in Compression: the Impact of Lossy Image Compression on Variable Size Object Detection within Infrared Imagery