Infima PLAN

The proprietary image compression format used by Infima Technology utilities is also termed PLAN. The PLAN format includes specialized models for processing and recompressing color .bmp, .tiff, and .jpeg files as well as raw medical images.

PLAN Lossless for JPEG was developed as a late addition to the Infima LAN format, using a completely different context modeling technique from the LAN. It uses a predictive scheme based on further compressed by partially uncompressing back to the DCT coefficients and provide context for the next Huffman code.

PLAN for Medical images was developed with the aim of Providing a low complexity lossless and "near lossless" image compression algorithm that could be able to offer better compression efficiency than lossless JPEG. The first EDition was finalized at the mid of 2007. The core of PLAN for Medical imaging is based on the LAN algorithm that relies on context modeling prediction integrated with neural network, as well as, residual modeling and context-based coding of the residuals. Most of the low complexity of this technique comes from the assumption that prediction residuals follow a two-sided geometric distribution (also called a discrete Laplace distribution) and from the use of Golomb-like codes, which are known to be approximately optimal for geometric distributions. Besides lossless compression, PLAN also provides a lossy mode where the maximum absolute error can be controlled by The encoder.

PLAN for JPEG2000 includes a lossless mode for JPEG2000 images based on a context adaptive nonlinear predictor that operates in either binary or continuous-tone modes, depending on the context of the current pixel. In the continuous-tone mode, gradient adjusted prediction takes place and is further improved by error feedback, where prediction errors are modeled under different contexts resulting in reduced conditional entropies. The coding step involves neural network predictions coding.