Perception based down sampling for low bit rate image coding

Abstract

Better image coding quality can be achieved by down sampling the visual content prior to compression and restore the full resolution content at the decoder when a higher compression rate is required. In this paper, a perception based adaptive sampling for low bit rate coding method is proposed. Sampling directions/ratios and the corresponding quantization parameters are determined adaptively, with the consideration of not only the local visual content but also the human attention model. Experiments show that the proposed method can yield more desirable compressed visual content than the other sampling based coding methods (and also in comparison with coding methods without down sampling) and automate the switching of the sampling modes for variable bit rate application. © 2009 Springer-Verlag Berlin Heidelberg.