Time-Space Weighting for Image Sequence Quantization

Abstract. This paper introduces a method for quantization of imagesequences which takes into account the human sensitivities in both space and time. A weighted clustering approach is used for quantization which allows exibility in the choice of weights. Assigning weights proportional to the space gradients and the time gradients is shown to produce better quantization of color image sequences. Image Analysis Applications and Computer Graphics, Proc.ICSC-95 p.389-396, Dec 1995

• Type: Conference
• Year: 1995
• Where: ICSC
• Authors: Hagit Zabrodsky Hel-Or
• Keywords: Better Quantization, Color Image Sequences, ICSC 1995, Image Analysis, Weighted Clustering Approach