Hiromi Yoshida

A New Binarization Method for a Sign Board Image with the Blanket Method

We propose a new binarization method suited for character extraction from a sign board in a scenery image. The binarization is one of a significant step in character extraction in order to get high quality result. Character region on sigh board, however, has many variation and colors. In addition to it, if there exists high frequency texture region like a mountain or trees in the background, it can be a cause of difficulty to binarize an image. At the high frequency region, the binarized result is sensitive to the threshold value. On the other hand, a character region of sign board consists of solid area, that is, includes few high frequency regions, and has relatively high contrast. So the binarized result of character region is stable at an interval of the threshold value. Focusing attention on this point, we propose a new binarization method which obtains a threshold value based on the fractal dimension by evaluating both regions density and stability to threshold value. Through the proposed method, we can get a fine quality binarized images, where the characters can be extracted correctly.

A Binarization Method for a Scenery Image with the Fractal Dimension

We propose a new binarization method suited for character extraction from a sign board in a scenery image. The binarization is thought to be a significant step in character extraction in order to get high quality result. Character region of sigh board, however, has many variation and colors. In addition to it, if there exists high frequency texture region like a mountain or trees in the background, it can be a cause of difficulty to binarize an image. At the high frequency region, the binarized result is sensitive to the threshold change. On the other hand, a character region of sign board consists of solid area, that is, includes few high frequency regions, and has relatively high contrast. So the binarized result of character region is stabile against an interval of the threshold value. Focusing attention on this point, we propose a new method which obtains a threshold value based on the fractal dimension to evaluate both regions density and stability to threshold change. Through the proposed method, we can get a fine quality binarized images, where the characters can be extracted correctly.

Automatic pencil sketch generation by using canny edges

This paper presents a system that automatically converts 2D raster images to sketch style. The proposed method first extracts edges at different resolutions. Then, these shapes and brightness are varied and merged. This process expresses trial and error in the actual sketch. Experimental results showed that the proposed method produces images with natural appearance.

Automatic draft reading based on image processing

Abstract. In marine transportation, a draft survey is a means to determine the quantity of bulk cargo. Automatic draft reading based on computer image processing has been proposed. However, the conventional draft mark segmentation may fail when the video sequence has many other regions than draft marks and a hull, and the estimated waterline is inherently higher than the true one. To solve these problems, we propose an automatic draft reading method that uses morphological operations to detect draft marks and estimate the waterline for every frame with Canny edge detection and a robust estimation. Moreover, we emulate surveyors’ draft reading process for getting the understanding of a shipper and a receiver. In an experiment in a towing tank, the draft reading error of the proposed method was <1  cm, showing the advantage of the proposed method. It is also shown that accurate draft reading has been achieved in a real-world scene.

Detection of view-disturbing noise by using time-axial clustering in spatio-temporal image

Lately, surveillance cameras are widely utilized. Nevertheless, raindrops or mud sometimes obstruct the field of view when we use them outside. In this research, we propose a method that can remove noises better than conventional method. In the conventional method, we take photographs while rotating the camera. Then, by considering the relationship between angles of photographing and objects in the images, we can detect noises. But in this method, if the luminance of the objects and noises are so close, we cannot detect them exactly. So we solve the problem by using clustering.