Video comperssion using multiwavelet and multistage vector quantization

Abstract EN

This paper presents a new video coning technique using multiwavelet transform and multi-stage vector quantization.

Three types of redundancies that are common in video sequence are spatial, temporal and psycho visual redundancies.

In this work, the spatial redundancy in the video is minimized using multiwavelet transfrom.

The transform coefficients are then quantized using multi-stage vector quantization scheme.

Motion estimation / compensation reduce temporal redundancy by exploiting interdictor correlation.

Kite cross diamond search is the block matching algorithm used for motion estimation.

The objective is to develop low bit rate video coder with acceptable visual quality.

The performance of the proposed method is compared with wavelet based video coder.