Masako Omachi

Traffic light detection with color and edge information

The concern of the intelligent transportation system rises and many driver support systems have been developed. In this paper, a fast method of detecting a traffic light in a scene image is proposed. By converting the color space from RGB to normalized RGB, some regions are selected as candidates of a traffic light. Then a method based on the Hough transform is applied to detect an exact region. Experimental results using images including a traffic light verifies the effectiveness of the proposed method.

Fast Template Matching With Polynomials

Template matching is widely used for many applications in image and signal processing. This paper proposes a novel template matching algorithm, called algebraic template matching. Given a template and an input image, algebraic template matching efficiently calculates similarities between the template and the partial images of the input image, for various widths and heights. The partial image most similar to the template image is detected from the input image for any location, width, and height. In the proposed algorithm, a polynomial that approximates the template image is used to match the input image instead of the template image. The proposed algorithm is effective especially when the width and height of the template image differ from the partial image to be matched. An algorithm using the Legendre polynomial is proposed for efficient approximation of the template image. This algorithm not only reduces computational costs, but also improves the quality of the approximated image. It is shown theoretically and experimentally that the computational cost of the proposed algorithm is much smaller than the existing methods.

Detection of traffic light using structural information

Driver support systems using images are paid attention and various researches on recognizing and understanding the road environment have been done. If it is possible to detect and recognize traffic lights, it will give useful information to a driver to understand the road environment. In this paper, a method of detecting a traffic light in a scene image is proposed. Considering the structure of a traffic light, we propose a method for detecting a traffic light based on the Hough transform. Experimental results using images including a traffic light taken by a digital camera through a windshield verifies the effectiveness of the proposed method.

An Approximation Method of the Quadratic Discriminant Function and Its Application to Estimation of High-Dimensional Distribution

In statistical pattern recognition, it is important to estimate the distribution of patterns precisely to achieve high recognition accuracy. In general, precise estimation of the parameters of the distribution requires a great number of sample patterns, especially when the feature vector obtained from the pattern is high-dimensional. For some pattern recognition problems, such as face recognition or character recognition, very high-dimensional feature vectors are necessary and there are always not enough sample patterns for estimating the parameters. In this paper, we focus on estimating the distribution of high-dimensional feature vectors with small number of sample patterns. First, we define a function, called simplified quadratic discriminant function (SQDF). SQDF can be estimated with small number of sample patterns and approximates the quadratic discriminant function (QDF). SQDF has fewer parameters and requires less computational time than QDF. The effectiveness of SQDF is confirmed by three types of experiments. Next, as an application of SQDF, we propose an algorithm for estimating the parameters of the normal mixture. The proposed algorithm is applied to face recognition and character recognition problems which require high-dimensional feature vectors.

Fast two-dimensional template matching with fixed aspect ratio based on polynomial approximation

Template matching is one of the fundamental techniques for signal and image processing. It has many applications such as detection, recognition, registration, retrieval, etc. One of the drawbacks of the template matching is the high computational complexity. In this paper, we focus on the two-dimensional image template matching with fixed aspect ratio and propose a method for speeding up the calculation. In the proposed method, a template is approximated by a polynomial in advance. Given an input image, the normalized cross correlations of the template and the regions of the input image are calculated efficiently with the polynomial. Experimental results using various sizes of images show the effectiveness of the proposed method.

Fast calculation of continuous wavelet transform using polynomial

This paper proposes an efficient calculation method of the continuous wavelet transform (CWT) at various scales. In the proposed method, the mother wavelet is represented by a polynomial. Then an efficient incremental algorithm for calculating the convolutions necessary for CWT using the polynomial is presented. Experimental results using two-dimensional data clarifies that the proposed method is much faster than the traditional methods.

Traffic Light Detection Considering Color Saturation Using In-Vehicle Stereo Camera

One of the major causes of traffic accidents according to the statistical report on traffic accidents in Japan is the disregard of traffic lights by drivers. It would be useful if driving support systems could detect and recognize traffic lights and give appropriate information to drivers. Although many studies on intelligent transportation systems have been conducted, the detection of traffic lights using images remains a difficult problem. This is because traffic lights are very small as compared to other objects and there are many objects similar to traffic lights in the road environment. In addition, the pixel colors of traffic lights are easily over-saturated, which renders traffic light detection using color information difficult. The rapid deployment of the new LED traffic lights has led to a new problem. Since LED lights blink at high frequency, if they are captured by a digital video camera, there are frames in which all the traffic lights appear to be turned off. It is impossible to detect traffic lights in these frames by searching the ordinary color of traffic lights. In this paper, we focus on the stable detection of traffic lights, even when they are blinking or when their colors are over-saturated. A method for detecting candidate traffic lights utilizing intensity information together with color information is proposed for handling over-saturated pixels. To exclude candidates that are not traffic lights efficiently, the sizes of the detected candidates are calculated using a stereo image. In addition, we introduce tracking with a Kalman filter to avoid incorrect detection and achieve stable detection of blinking lights. The experimental results using video sequences taken by an in-vehicle stereo camera verify the efficacy of the proposed approaches.

Element-Level Clustering of Feature Vectors Considering Correlations for Analyzing Image Data

Clustering is a fundamental tool for data analysis. Typically, all attributes of the data are used for clustering. However, if a set of attributes can be divided into meaningful subsets, it may be effective to cluster the data for each subset. In this paper, we propose a method for dividing the set of elements of feature vectors into meaningful subsets. Considering the dependencies between the elements, the correlation is used as the metric for clustering. In order to effectively solve the optimization problem, a technique for graph cut is used. After dividing the set of elements into subsets, clustering is performed for each subset. Experiments using a handwritten image database show the effectiveness of the proposed method.

Pattern recognition using boundary data of component distributions

In statistical pattern recognition, a Gaussian mixture model is sometimes used for representing the distribution of vectors. The parameters of the Gaussian mixture model are usually estimated from given sample data by the expectation maximization algorithm. However, when the number of data attributes is large, the parameters cannot be estimated correctly. In this paper, we propose a novel approach for estimating the parameters of the Gaussian mixture model by using sample data located on the boundary of regions defined by the component density functions. Experiments are carried out to show the characteristics of the proposed method.