In this paper we examin the application of self-organizing list to vector quantizer for digital image coding. Owing to computational complexity and memory storage , Vector quantizer can not be able to remove redundancy entirely in image coding , and transform image into correlated indexes. This correlated indexes do not span distance space , so can not data compressing using distance based coding schemes. On the other hand , in point of data compression of simbols (or words) , self-organizing list coding scheme has been used. We intende to remove redundancy of indexes applying self-organizing List to vector quantizer. In this paper, self-organizing list coding of vector-quantized indexes of image, and new vector quantization sccheme which has self-organizing structure codebook are investigated. And their simulation results using monochrom images show advantage of application of self-organizing list.
A realtime displaying method of colored stereographic solid images following the movement of visual point is presented. CRT on the rotatable rack is turned automatically toward the visual point and the image corresponding to the point is immediately displayed. The chain model is proposed to represent the shape of solid. This model enables the table look-up operation of coordinate transformation and hidden surfase removal. Terefore. 5M voxels/sec of operating rate is realized for the realtime operation using widely used memories and simple control circuits.
The authors have already presented Structure Runlength Coding, which is referred as SRC. SRC is a data compression coding scheme for digital pictures. It gives completely reversible codesfor any given pictures. It works well especally on multi-colored CG pictures. After further study on SRC, some redundancy has been found in the way of code generation. This report introduces an improved coding method employing a two-dimensional color-predictor in the placeof pixel reference mechanisms of the conventional SRC. Some results of experiments are given.
In this paper, we report an extraction method of contour lines from binary image data picked up from a map of I to 25.000. By the method contour lines are extracted and given the altitude. This algorithm can connect sprit coutour lines. Next we describe the result of encoding the extraction data to 3-directed optimized chain code. This code can compress the data to 1.5 bit / pixel.
In this paper, a fast method is described which carries out geometrical transformation such as translation, enlargement, reduction and rotation. The purpose is real time transformation at low cost, low power and small size hardware. Address control memory plays a very important part in this.
This paper descrlbes transformation errors produced by Image transformations using address control memory. The errors have some regularity and closer relations with the distance from an integer. Computer simulation of the transformations is made, where address control memory has 0〜8 bits decimal memory.
This paper is a report on the simulation system of a robot manipulator using a personal computer. The interpolation is used for the smoothing representation of manipulator movements as the moving speed changes. A system using DSP(Digital Signal Processor) for fast image processing is made for trial. It is achieved that the computing time is reduced to 1/110 - 1/170, and the whole processing time to 1/5 - 1/8 compared with the system without DSP. As a result, it is con-firmed that our system is effective on the manipulator simulation.