Naslov
Lossless Image Compression Based on Contextual Adaptation of Predictor Blends

Autori
Knezović, Josip ; Žagar, Martin ; Kovač, Mario

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
First Croatian Computer Vision Workshop 2012. Abstract Book / - Zagreb : Center of Excellence for Computer Vision, University of Zagreb, 2012, 11-11

Skup
First Croatian Computer Vision Workshop

Mjesto i datum
Zagreb, Hrvatska, 2012

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Nije recenziran

Ključne riječi
lossless image compression; image coding; predictive coding; predictor blends; statistical coding

Sažetak
Lossless compression of visual data is required in applications such as medical imaging, image archival, remote and satellite imaging etc. We present a new adaptive predictive image coding method based on the blending of multiple static predictors on a dynamically classified causal context of neighboring pixels. The idea of predictor blends is further expanded through the determination of blending context that changes its shape on a pixel--by--pixel basis using simple classification technique, thus allowing the modeling more complex image structures such as nontrivially oriented edges, the periodicity and coarseness of textures. Typical natural images are characterized as being composed of image regions with different local properties. The predictor estimates those properties around the currently unknown pixel and adjusts itself so that the presence of detected properties affect the way the final prediction is made. Predictive part is followed by a heuristic contextual model and statistical encoder.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo

POVEZANOST RADA