images.txt

# JPEG
public
lossly compression (different compression levels)

# PNG
public
lossless compression
vector graphics

# GIF
private (patented)
lossless compression

# TIFF
raster graphics (pixels, rgb)
pushed by Adobe
supports different compressions: lossy like jpeg and lossless less lzw
supports storing multiple images
* as single slice, which can be organized in tree
* does not support storing 3D tensor -> DICOM better
Supports different color palettes: grayscale, rgb, cmyk
Metadata (proprietary infos) can be stored as private tags

# DICOM
Standard for communicating medical data, in particular images
Organizes as datasets
Each datasets can have multiple attributes
  * good for storing meta info, e.g. about patient
One attribute can be nd tensor of pixels
Images can be compressed by different lossy and lossless compression methods
pydicom library
3DSlicer for visualization