Image Coding Based on Second Generation Bandelet Transform

Bandelet lead to optimal sparse representations for geometrically regular images,and have great potential in imge compression.This paper describes a new image coding strategy based on the second generation Bandelet transform.Bandelet coefficients are first quantized with a nearly uniform quantizer.With the resetting of Bandelet coefficients,a context model of Bandelet coefficients for image compression is proposed,and then a compression scheme is derived.Coefficients of the lowest subband are coded with CALIC model.Experimental results show that the model is effective in image compression.For images with regular geometric flow,the compression scheme achieves higher quality than JPEG-2000.