Norio Ito

Temporal-scalable coding based on image content

An object-based temporal scalability codec is proposed by introducing shape coding, a new motion estimation/compensation method, weighting techniques, and background composition. The major feature of this technique is determining the frame rate of the selected objects in the motion picture individually so that the motion of the selected region is smoother than that of the other area. The observation of the computer simulation proves that the proposed method achieves the better image quality and it enables us to represents the motion of the selected objects hierarchically.

Object wavelet transform for coding of arbitrarily shaped image segments

Object-based coding is one of the most challenging key features which the ISO/IEC MPEG-4 is trying to standardize. One approach to transform an arbitrary shaped image region is addressed. The proposed approach called object-based wavelet transform (OWT), is simple to implement and is a smooth extension of regular wavelet transform (WT). The OWT consist of two phase processes: one is an extrapolation phase for regular WT and the other is a coefficients handling phase for eliminating redundancy caused by the extrapolation. Due to some experimental results, it is confirmed that the method performs the same as other shape-adaptive approaches with low complexity.