Visual data rate gain for wavelet foveated image coding

Wavelet image coding exhibits a robust error resilience performance by utilizing the naturally layered bitstream construction over a band-limited channel. In this paper, a closed form of the visual entropy is defined based on a visual weight over the wavelet domain, which is characterized by the HVS (human vision system) over the frequency and spatial domains. This visual weight is then utilized as a criterion for determining the coding order of wavelet coefficients and the resulting improved visual quality is demonstrated. In terms of visual entropy, a transmission gain of up to 20 % can be obtained at a normalized channel throughput of about 0.3.