Makoto Miyahara

Mathematical Transform Of (R, G, B) Color Data To Munsell (H, V, C) Color Data

In studying new-generation color image codings, it is very effective 1) to code signals in the space of inherent tri-attributes of human color perception, and 2) to relate a coding error with perceptual degree of deteriorations. For these purpose, we have adopted the Munsell Renotation System in which color signals of tri-attributes of human color perception (Hue, Value and Chroma) and psychometrical color differences are defined. In the Munsell Renotation System, however, intertransformation between (RGB) data and corresponding color data is very cumbersome. Because the intertransformation depends on a look up table. This article presents a new method of mathematical transformation. The mathematical transformation is obtained by multiple regression analysis of 250 color samples, which are uniformly sampled from whole color ranges that a conventional NTSC color TV camera can present. The new method can transform (RGB) data to the data of the Munsell Renotation System far better than the conventional method given by the CIE(1976)L*a*b*.

Block Distortion in Orthogonal Transform Coding--Analysis, Minimization, and Distortion Measure

Psychophysical considerations show that the block-shaped distortion peculiar to orthogonal transform coding (OTC) is ten times more objectionable than random noise distortion. Minimizing the blockshaped distortion is considered here by analyzing the process of generation of coding errors, and the required characteristics of the orthogonal transform function (OTF) are clarified. Computer simulations substantiate its validity. Graphical illustration and measures of the block-shaped distortion are proposed and substantiated. With the aid of these measures, one can evaluate the performance of OTC or a new OTF, if proposed, without conducting visual assessment tests.

Objective picture quality scale for video coding

To design the image communication system including human visual system systematically, the objective quality estimation method of the picture based on the model of the human vision is necessary. We develop the objective picture quality scale (PQSvideo) for video coding considering the relation between the physical distortion factors and the psychological picture estimation factors. The obtained cross correlation coefficient between the PQSvideo and the mean opinion score (MOS), becomes 0.978.

Generalized Karhunen-Loeve Transformation I (Theoretical Consideration)

A generalized approach to obtain a family of the reversible linear transform functional sets for the transform compression is proposed in this work. The present method is formulated so as to include the conventional Karhunen-Loeve (KL) transform method in itself, and shows how one might take account of a spatially inhomogeneous and correlated error measure. These new functional sets can be derived minimizing a generalized measure defined in a skew linear error space instead of the conventional orthogonal one. The generalized error measure is minimized to obtain an optimal functional set which is capable of making the errors spatially inhomogeneous and reducing the correlation between them. For a strongly correlated input data sequence, an algebraic equation for the sinusoidal inhomogeneity of errors is analytically solved and found to yield the discrete Mathieu functions (DMFs) expanded m terms of the discrete cosine functions (DCFs). As an example, the simulation results for the DMFs are shown in comparison with the performances of the KL and the DC transform compressions. Since, as is well known, the spatially, highly correlated, and localized errors are very visible to the human eye, one may expect that the present method is effectively applicable to the block transform compression of image data to reduce the visible distortion.

Region segmentation using K-mean clustering and genetic algorithms

One of the hard problems in image recognition and understanding is region segmentation. A traditional segmentation method such as clustering is not fully useful for any image, because of the initial values of clusters and the evaluation functions of segmented clusters affect the results of region segmentation. To solve this problem, we introduce the genetic algorithm (GA) for clustering. The experimental result shows the satiable results of region segmentation which have been achieved by applying GA.<<ETX>>

Image segmentation based on ULCS color difference

In studying high efficiency color image coding and image processing, it is important to segment several regions that represent real objects. By performing this region segmentation, we can establish the structural description using the characteristic information of regions. To segment the image into several characteristic regions, we adopt the clustering algorithm in the Uniform Lightness Chromaticness Scale System and the merging process based on the measure of Godloves color difference.

Objective picture quality scale by neural network fed the distortion factors to the input layers

It is important to improve the picture quality of an image reconstructed by computer simulation. It is also important that the performance of coders be evaluated. To this end, an objective picture quality scale is essential for image processing. This paper analyzes human perceptual properties in detail, formulizes approximately the fundamental image disturbance factors and proposes an objective picture quality scale (PQS). As a test image, a coding error image was fed into a neural network and experiments were conducted with a method to evaluate the picture quality. Finally, the new objective picture quality scale obtained herein can estimate the mean opinion scores (MOS) of two kinds of previously untaught images whose characteristics differ greatly.

Colour image coding based on Munsell specification system.

人間の色彩感覚を適確に表現し, かつその尺度を均等に表現する表色系であるマンセル (HVC) 表色系に注目し, カラー画像の伝送時に生ずるひずみを色知覚の許容限以下にするような色立体分割方法を考案した.また, 量子化された色情報の代表値を予測符号化する伝送方式を検討した.この結果, エントロピー約3.5 (bit/pixel) で, ほぼ満足できる画質の画像を伝送できた.色感覚をあらわすH, V, Cは互いの独立性が良いので, Y, I, Qと比較した場合, 色情報を無駄なく符号化することができ, また輪郭部の情報損失も少なくできるなど, 多くの利点がある.