The Department of Image Processing, Institute of Information Theory and Automation, Czech Academy of Sciences, engaged in image processing and pattern recognition research. Their primary application areas include biomedicine, industry, remote sensing, astronomy, and art conservation.

The department's research encompasses

• the recognition of distorted images and patterns using invariant descriptors

• the registration and fusion of multiple images for enhanced quality

• the development of the theory of moment invariants like rotation, affine, and convolution invariants

• the restoration of degraded images through techniques such as multichannel blind deconvolution, edge-preserving denoising, local contrast enhancement, and color transformations.

• Methods based on neural networks, particularly focusing on equivariant networks