Region of interest coding for low bit rate image transmission
ABSTRACT Region of interest (ROI) coding has become a new challenge for low
bandwidth media channels. This paper presents an original method for low
bit rate compression using this concept. It is based on a new
progressive coding technique called LAR (Locally Adaptive Resolution)
combining a spatial nonuniform sampling and a variable block-size
spectral transform. Only little overhead is required for the ROI shape
description, and the data structure enabling the inner ROI enhancement
is totally correlated to this shape. This technique is also very fast,
allowing real-time and interactive multimedia applications
- SourceAvailable from: Samir Amir
- "Inspiré par cette constatation, un nouveau schéma d'un codage vidéo scalable basé sur une approche pyramidale de compression d'image fixe (LAR-APP) a ´ etéélaboré . La méthode LAR (Locally Adaptive Resolution) fondée sur une représentation de l'imagè a taille de bloc variable définit une technique efficace de compression sans perte . Son principe de codage a inspiré trois approches pyramidales de compression d'images fixes, le LAR-APP, l'Interleaved S+P et le RWHT+P   . "
Conference Paper: LAR vidéo: Codage sans perte à scalabilité sémantiqueCORESA 2006; 09/2006
[Show abstract] [Hide abstract]
- "In order to meet the requirement, a new scalable video coding algorithm, based on an efficient multiresolution approach for still image compression (LAR-APP) has been developped. The LAR (Locally Adaptive Resolution) based on a variable block-size decomposition, leads to an efficient lossy image compression technique . Wrought on this coding assumption , three LAR-like pyramidal decomposition techniques namely LAR-APP , Interleaved S+P  and RWHT+P  followed it. "
ABSTRACT: Baptized "LAR Video", method proposed in this paper describes a new lossless video coding algorithm with ad-vanced semantic scalability. Motion estimation and compensation steps are first achieved to produce the well known displaced frame difference (DFD). The basic idea is to apply on this residual error a pyramidal decomposition based on an efficient scalable image compression technique, named LAR-APP. Resulting from a progressive data broadcasting, image-sequence can be scalably rebuilt in the decoder at different spatial resolution level. The given experimental results show that the proposed solution, in ad-dition to the scalability, achieves good compression performances.
[Show abstract] [Hide abstract]
- "In some previous works, we have investigated some coding schemes based on variable block size representations leading to efficient compression at both low and high bit rates . "
ABSTRACT: This paper presents a full multiresolution lossless coding method, with advanced semantic scalability. In particular, a reversible form of the usual Walsh Hadamard Transform (RWHT) is first introduced as an alternative to standard loss-less transform. A pyramidal representation and decompo-sition schemes involving this basic transform are then pro-posed. Significant improvements are obtained using two ad-ditional concepts: the "locally adaptive resolution" through a quadtree representation and a prediction step. The given ex-perimental results show that the proposed RWHT+P achieves excellent performances compared to state-of-the-art.