When the hybrid image coding method using motion-compensated prediction and transform is applied to very low bit-rate (less than 64kbit/s) image coding, it often causes block distortion in decoded images, since motion detection and motion-compensated prediction has been performed based on square block. Recently, as the one of the methods to solve this problem, some motion-compensated predictive methods based on affine transform or bilinear transform were developed. However, these methods can't exactly express the appearance motion in the 2-D projective image plane corresponding to the motion in the 3-D space. In this paper, we propose a new motion-compensated predictive method for very low bit-rate image coding. The proposed method is based on the perspective projection and the gradient of pixel value in time and space domain, and has high ability in motion representation. The computer simulation was performed using moving test images, and the resulting predictive image and SNR were compared with the conventional methods. The results showed that the proposed method has good performance as a motion-compensated predictive method.
In this paper, we propose a new half-pel motion vector search algorithm, which reduces the required computational power compared with the ordinary. two step algorithm. The proposed algorithm also consists of two steps. In the first step, it searches for motion vectors with integer-pel accuracy. using subsampled video data, and gets a set of motion vector candidates. In the second step, it, searches for the half-pel vector around the candidates using original (not subsampled) video data. The simulation results show that, the proposed algorithm is effective to reduce the number of operators without the loss of performance.
A motion compensated TV standards converter is newly. proposed. In order to estimate motion vectors suitable for field interpolation necessary for field rate conversion, 4 key techniques are utilized, which are prefiltering of input TV signals, improved gradient method applied to hierarchical two-stage motion estimation, post-processing of motion vectors, and digital filtering with many coefficients for line interpolation. From computer simulation experiments, significant, improvement of picture quality is expected, compared to conventional schemes with motion estimation.
The approximation method of arithmetic coding with multiplication of prime numbers Is descrlbed in this paper. This method makes it possible that the multiple operation for augend is transformed to additional operation of powers of the prime number. And the value of augend is stored in a register with high accuracy. Furthermore, we show the result of the simulation of Bi-level image coding using pr6 sented method.
Mulimedia applications has lately attracted wide attention, owing to the international standardisation of video coding algorithm (ISO/MPEG). On the other hand, server-client computing environment is rapidly spread in the areas of business, education, and reserchIn this paper, an experimental video on demand system based on Ethernet LAN is introduced. Inthis system, previously developped MPEG-1 real-time CODEC and decoder cards are used forauthoring system and client system respectively.
An optimum bit rate allocation method for multi-encoder system to achieve heigher performance is proposed. First we analyze the relation between bitrate and coding distortion, showing that the relation is well approximated by 1st order functions on a logarithmic scale. Then we propose a new bitrate allocation algorithms that minimize total coding distortion under the constraint of constant total bitrate and show that total visually weighted mean square error is always reduced.
There has been much interest in digital video applications, Supported by rapid progress of video coding technology and international standardization. ATM network is another key element to realize high quality audiovisual communication systems, which is under active standardization as well. This paper presents the current status and major outcome of the standardization work in ISO/IEC and ITU-T addressing multimedia multiplexing/demultiplexing.
In distributed multimedia systems, the concepts of quality of services (QOS) for continuous media data such as digital video and audio will become important. We have been developping Real-Time Mach for distributed multimedia computing, which is a real-time microkernel supporting these concepts. This paper presents several issues of operating system supports for continuous media.