Chie Muraki Asano

Extraction of trabecular structures of mandible excluding tooth roots on dental panoramic radiographs using mathematical morphology

In this paper, an extraction method of trabecular structures from dental panoramic radiographs using mathematical morphological operations is proposed. It can extract trabeculae excluding roots of teeth and enables the assessment of tooth extraction for trabecular pattern. A measurement method of the lengths and directions of trabecular segments is also proposed in this paper. It is suggested that the measurements of healthy and osteoporotic examples support our hypothesis that the trabeculae parallel to the roots are reduced more than those perpendicular to the roots by osteoporosis

Visual complexity assessment of painting images

In this paper, we propose a framework to assess visual complexity of paintings. This framework provides a machine learning scheme for investigating the relationship between human visual complexity perception and low-level image features. Since the global and local characteristics of paintings affect humans holistic impression and detail perception, we design a set of methods to extract the features that represent the global and local characteristics of paintings. By feature selection, we look into the role that each image feature plays in assessing visual complexity. Then the selected features are combined by a Support Vector Machine for classification. Experimental results indicate that the proposed work can predict the visual complexity perception of paintings with the accuracy of 88.13%, which is highly close to the assessments given by humans. Compared with the conventional measure of complexity, our approach considers human visual perception and performs more efficiently in assessing visual complexity of painting images.

Statistical quantification of the effects of viewing distance on texture perception

In general, viewers are more attracted to local features in images at a shorter viewing distance and to global features in images at a longer viewing distance. However, numerical analysis of the effect of viewing distance on human texture perception and how the perception of global and local changes under certain conditions are still undetermined. In this paper, we present statistical prediction of the relationship between the domination ratio of global and local features and the viewing distances under the control of several factors, using the logistic regression model. We synthesized textures by separately controlling global and local textural features using a texture model based on mathematical morphology, namely the primitive, grain, and point configuration texture model. Visual sensory tests were carried out on 80 subjects during two sets of experiments. The collected data were statistically analyzed using logistic regression and Akaike information criteria. Besides the main factor of viewing distance, the factors including gender, changing the order of viewing positions, and prior knowledge were also shown quantitatively to have significant influence on human texture perception. Our results showed that (1) local features of a texture were more attractive to females than males, (2) the first impression might have affected subsequent decisions in texture perception, and (3) subjects who had prior knowledge (supervised) were more sensitive to the changes in global and local dominance. (4) Regarding the interactions of the factors, prior knowledge reduced the effects of individual differences and perception condition differences on human texture perception. This study is dedicated to the construction of numerical relationships between viewing distance and human texture perception as well as to cognitive investigation of biases in global and local perceptions.

Primitive and grain estimation using flexible magnification for a morphological texture model

An improved morphological estimation method of textural elements based on the Primitive, Grain, and Point Configuration (PGPC) texture model is proposed. The PGPC texture model has shown promising applications such as noise removal, texture modification, and texture synthesis. However, the estimation is not always successful since the magnification process of the primitive to ensure the assumption that the grains are homothetic does not always fit to each image. We propose in this paper a novel estimation method introducing more flexibility into homotheticitys conventional assumption of the grains, and exploring a suitable structuring element for the homothetic magnification process of the primitive. Experimental results show that the proposed method provides more representative grains than the conventional method

Visual Complexity Perception and Texture Image Characteristics

Visual complexity perception is an important issue in the fields of psychology and computer vision because it leads to the better understanding of the nature of human perception as well as the properties of the objects being perceived. In this study, five important characteristics of texture images that affect visual complexity perception are identified: regularity, understandability, roughness, directionality, and density. Among these, understandability is a deterministic characteristic, which reflects the viewers prior knowledge and experience. These characteristics significantly affect the visual complexity perception of texture images. In order to achieve our objective, we carried out two experiments involving visual complexity assessment and paired comparison evaluation with 30 respondents. We applied correlation analysis, factor analysis, and multidimensional scaling to analyze the collected data. The experimental results showed that most of the human impressions of visual complexity can be explained by the perceived characteristics of texture images.

Evaluation of Human Visual Impressions in Gray Scale Textures Using Morphological Manipulation

An evaluation method of human visual impressions in gray scale textures using morphological morphology is proposed. Variations of textures are generated by modifying repetitively arranged objects and configurations of the arrangements of original textures. The variations are presented to human respondents, and similarity of modified textures based on human impressions is evaluated. The results of the human evaluation are compared with the results of similarity evaluation based on image features. It shows that global features such as density, regularity and directionality of the point configurations have significant effects on human visual impressions and identification of textures. In case of a texture without significant characteristics on its point configuration, local features such as the shape of the objects have some effects on the visual impressions.

Dual primitive estimation for morphological texture analysis

A method of estimating dual primitives in a textural image is proposed. This method is based on the Primitive, Grain, and Point Configuration (PGPC) texture model, which regards a texture as an arrangement of grains derived from one or a few primitives. Appropriate primitives can be represented by morphological structuring elements estimated from a texture. Conventional primitive estimation methods estimate only one primitive from each textural image. However, they do not work well on textural images that contain more than one basic structure, since two or more types of grain cannot be generated from only one primitive. The proposed method simultaneously estimates two optimal structuring elements of a texture. The experimental results show that the proposed method provides more representative estimations than the conventional method.

Application of Size Distribution Analysis to Wrinkle Evaluation of Textile Materials

An evaluation method of wrinkle shapes on fabrics using simple morphological operations is proposed. It calculates the size density function of the object indicating the region surrounded by a folded fabric in the experimental condition of the standard wrinkle/crease angle test. The characteristics of the size density function and the parameters of the linear and cubic function fitted to the size density function indicate the fabric shape characteristics, which correspond to visual edge sharpness and roundedness of pleat lines, as well as their mechanical properties.

An experimental study of the hybridization of logistic discriminant analysis and multilayer neural network for image identification

A hybridized classification system of the logistic discriminant analysis and the three-layer neural network is proposed. This system is basically a linear discrimination and is assisted by the neural network only for the cases that are difficult to be classified by linear methods. This system presents a simple discrimination structure given by linear methods, and its computational cost is much lower than the exclusive use of the neural network while the misclassification rate is as low as the neural network. The ability of this system is shown experimentally in the case of applying it to image identification problems. The computation time for the learning process is reduced to one-fifth by this method in this experiment, while the misclassification rate remains almost the same.