All) Get software for JPEG, JPEG2000, JPEG XR, JPEG XL and use it for compressing images to a fixed file size (compression ratio).

All) Get software for full-reference and blind (i.e. no-reference) quality metrics and learn to apply it and understand what it does.


I Double Compression

apply two different compression schemes one after the other, both with a fixed compression rate, e.g. 5; investigate good and poor
combinations wrt. quality - 2 groups max.

additional codecs: AVIF, BPG, WebP, HEIC / HEIF

II Recognition of Compression type

Compress testdata to (a) fixed filesize or (b) fixed quality and try to train a classifier to identify the compression scheme that has been used.

Then compress with JPEG and investigate how long (for decreasing JPEG quality) you can discriminate the different formats.

2 groups max

III Find software for Deep - learning based Lossy compression (JPEG AI) and learn to apply it (again: fixed target file size).

In particular: learn to train it, test on classes of similar biometric data as compared to common data (FPs, vascular data, microscopy data, etc.)
Comparison to non-learning based schemes - 2 different ones, 2 groups max

All) Prepare a presentation and give it