Yousun Kang

Texture Classification Using Hierarchical Linear Discriminant Space

As a representative of the linear discriminant analysis, the Fisher method is most widely used in practice and it is very effective in two-class classification. However, when it is expanded to a multi-class classification problem, the precision of its discrimination may become worse. A main reason is an occurrence of overlapped distributions on the discriminant space built by Fisher criterion. In order to take such overlaps among classes into consideration, our approach builds a new discriminant space by hierarchically classifying the overlapped classes. In this paper, we propose a new hierarchical discriminant analysis for texture classification. We divide the discriminant space into subspaces by recursively grouping the overlapped classes. In the experiment, texture images from many classes are classified based on the proposed method. We show the outstanding result compared with the conventional Fisher method.

Semantic Segmentation and Object Recognition Using Scene-Context Scale

Scene-context plays an important role in scene analysis and object recognition. Among various sources of scene-context, we focus on scene-context scale, which means the effective region size of local context to classify an image pixel in a scene. This paper presents semantic segmentation and object recognition using scene-context scale. The scene-context scale can be estimated by the entropy of the leaf node in multi-scale text on forests. The multi-scale text on forests efficiently provide both hierarchical clustering into semantic textons and local classification depending on different scale levels. For semantic segmentation, we combine the classified category distributions of scene-context scale with the bag-of-textons model. In our experiments, we use MSRC21 segmentation dataset to assess our segmentation algorithm and show that the usage of the scene-context scale improves recognition performance.

Depth Perception from a 2D Natural Scene Using Scale Variation of Texture Patterns

In this paper, we introduce a new method for depth perception from a 2D natural scene using scale variation of patterns. As the surface from a 2D scene gets farther away from us, the texture appears finer and smoother. Texture gradient is one of the monocular depth cues which can be represented by gradual scale variations of textured patterns. To extract feature vectors from textured patterns, higher order local autocorrelation functions are utilized at each scale step. The hierarchical linear discriminant analysis is employed to classify the scale rate of the feature vector which can be divided into subspaces by recursively grouping the overlapped classes. In the experiment, relative depth perception of 2D natural scenes is performed on the proposed method and it is expected to play an important role in natural scene analysis.

Texture Structure Classification and Depth Estimation using Multi-Scale Local Autocorrelation Features

While some image textures can be changed with scale, others cannot. We focus on a multi-scale features of determing the sensitivity of the texture intensity to change. This paper presents a new method of texture structure classification and depth estimation using multi-scale features extracted from a higher order of the local autocorrelation functions. Multi-scale features consist of the meansand variances of distributions, which are extracted from theautocorrelation feature vectors according to multi-level scale. In order to reduce dimensional feature vectors, we employ the Principal Component Analysis (PCA) in the autocorrelation feature space. Each training image texture makes its own multi-scale model in a reduced PCA feature space, and the test of the texture image is projected in the homogeneous PCA space of the training data. The experimental results show that the proposed multi-scale feature can be utilized notonly for texture classification, but also depth estimation in two dimensional images with texture gradients.

Scale-Optimized Textons for Image Categorization and Segmentation

Texton is a representative dense visual word and it has proven its effectiveness in categorizing materials as well as generic object classes. Despite its success and popularity, no prior work has tackled the problem of its scale optimization for a given image data and associated object category. We propose scale-optimized textons to learn the best scale for each object in a scene, and incorporate them into image categorization and segmentation. Our textonization process produces a scale-optimized codebook of visual words. We approach the scale-optimization problem of textons by using the scene-context scale in each image, which is the effective scale of local context to classify an image pixel in a scene. We perform the textonization process using the randomized decision forest which is a powerful tool with high computational efficiency in vision applications. Our experiments using MSRC and VOC 2007 segmentation dataset show that our scale-optimized textons improve the performance of image categorization and segmentation.

A method for detecting human face region based on generation and selection of kernel features

Recent researches for detecting face regions from images have paid attention to high dimensional kernel features (KFs), which are obtained by a non-linear transformation of original features extracted from images. A support vector machine (SVM) is one of the most prominent learning algorithms for KFs. However, SVM is time-consuming because of needing a large number of KFs to improve the accuracy of the classification. This paper proposes a new method that constructs a classifier between face and non-face regions by generating and choosing KFs based on Kullback-Leibler divergence (KLD). The KLD means a distance between two distributions of face and non-face data under a given KF, and some KFs of large KLDs are selected for the face detection. Moreover, the use of KLD enables us to generate new KFs. and to deal with different kinds of KFs concurrently. Some experiments show that our method can reduce the number of KFs much more than SVM, and achieve almost equal or better detection rate than that of SVM.

Texture classification using hierarchical discriminant analysis

As the representative of the linear discriminant analysis, the Fisher method is most widely used in practice and it is very effective in two-class classification. However, when it is expanded to multi-class classification problem, the precision of its discrimination may become worse. One of the main reasons is an occurrence of overlapped distributions on a discriminant space built by Fisher criterion. In order to take such overlap among classes into consideration, our approach builds a new discriminant space with hierarchical tree structure for overlapped classes. In this paper, we propose a new hierarchical discriminant analysis for texture classification. We can divide a discriminant space into subspace by recursively grouping overlapped classes. In the experiment, texture images of many classes are classified based on the proposed method, and we show the outstanding result compared with the conventional method.