A relative quality controlled region-of-interest image coding based on wavelet transform

A region-of-interest (ROT) image coding based on wavelet transform that can control the image quality between ROT and region-of-non-interesting (non-ROI) under a constrained quantity of information is proposed in this paper. We introduce a scalar parameter referred as β to control, and β is ratio of the quality ROT to non-ROI and is relative measure. Assignment of the total quantity of information, i.e. bit-rate, to each region can be derived analytically by the method of Lagrange multipliers. A method which can reduce the information quantity of the shape of ROI/nonROI to be transmitted to the receive end as a side-information will be also proposed. Our method can avoid so many information loss for nonROI at lower constrained bit-rate caused in JPEG-2000, therefore the subjective image quality for whole of restored image will be improved, maintaining the quality of ROT is higher than the non-ROI.