1) JPEG, JPEG2000, JPEG XR, JPEG XL, JPEG AI compression to a fixed file size; evaluation with full references image quality metrics (IQM) beyond PSNR and SSIM variants

2) JPEG, JPEG2000, JPEG XR, JPEG XL, JPEG AI compression to a fixed file size; evaluation with no reference image quality metrics (IQM)

3) Recognition of compression type (JPEG, JPEG2000, JPEG XR, JPEG XL, JPEG AI, Compress AI): Compress testdata to fixed _filesize_  and train a classifier to identify the compression scheme that has been used. Additional codecs: AVIF, BPG 

4) Recognition of compression type (JPEG, JPEG2000, JPEG XR, JPEG XL, JPEG AI, Compress AI): Compress testdata to fixed _quality_ and train a classifier to identify the compression scheme that has been used. Additional codecs: WebP, HEIC / HEIF

5) Beyond JPEG AI and Compress AI: more recent compression techniques - do we get even better ?

6) Re-train JPEG AI/Compress AI (or other learning-based codecs) to specific data: Microscopy cell segmentation - besides IQM, evaluate impact on segmentation

7) Re-train JPEG AI/Compress AI (or other learning-based codecs) to specific data: Vascular biometric data - besides IQM, evauate impact on vascular recognition

8) Re-train JPEG AI/Compress AI (or other learning-based codecs) to specific data: Iris Data - besides IQM, evaluate impact on iris segmentation and recognition

All) Reference quality metrics: PSNR, SSIM

All) Experimental datasets need to be uncompressed or compressed in lossless manner (e.g. png)

All) Prepare a presentation and give it at the end of the term - WHEN ??