Akio Okazaki

An automatic circuit diagram reader with loop-structure-based symbol recognition

A high-performance logic circuit diagram reader was developed for VLSI-CAD data input. Almost all logic circuit symbols include one or more loop structures. A description is given of an efficient method for recognition of these loop-structured symbols. The proposed method consists of two processes: symbol segmentation and symbol identification. Symbol identification is achieved by a powerful hybrid method which uses heuristics to mediate between template matching and feature extraction. The entire symbol recognition process is carried out under a decision-tree control strategy. The entire recognition system for circuit diagrams is briefly explained, including character string recognition and connecting line analysis. >

High quality vectorization based on a generic object model

We propose a new method for high quality vectorization based on a generic object model. In conventional methods, the quality of vectorization has been regarded as the same thing as digitizing accuracy, which is evaluated in terms of the error between the original line-drawing and the resultant vector sequence. However, this accuracy does not always correspond to perceptual quality. Based on the discussion about possible distortions in a vectorization process which strongly affects perceptual quality, we introduce a generic model of an object. The generic properties are described in the object model, for example, “The object is a polygon whose corners all have a right angle” or “The object is composed of several pairs of almost parallel lines”. The method consists of three processes: pre-vectorization, object recognition and shape modification based on an object model. The approximation line-figure which is first defined in the pre-vectorization process, is modified so as to meet the properties described in the object model. A cost function is derived from the model so that shape modification can be formalized as an optimization problem. A relaxation method is employed to save on computation time. Object recognition is needed for object model selection. The method was applied to geographic maps for urban planning to extract building polygons. Subjective evaluation on the extracted building shapes showed that 93.6% of the buildings were vectorized with high quality and that the number of buildings with insufficient quality was reduced to one twelfth.

Detecting intruders using time-series data by projection pattern of silhouette

This paper describes a method of detecting intruders by video images processing. There are some studies on detecting intruders in video image by analyzing difference-binary images and motion vectors. Because of the use of simple features such as size and motion vectors of changing regions, these methods are not fully able to reduce erroneous detections caused by other factors, such as small animals, swaying of trees, and changes in brightness. It is proposed a method of detecting intruders by time-series data on the projection pattern of the silhouette. In this method, changing regions in video images are extracted by the interframe differential technique and are transformed into projection patterns. Evaluation scores (a measure of similarity to a human silhouette) are calculated by flexible matching between observed patterns and the standard pattern. Discrimination processing is carried out in terms of a time-series score by tracking a target of each video frame. Experiments have been made with a prototype and confirmed the effectiveness of this method.

Access Control System with Automatic Logging for User's Facial Information

An access control system that automatically logs in a user by his or her facial information is described. Furthermore, a method for automatically tracking logged data, based on a repetitive operation of template improvements, is shown. Face recognition is a favorable variety of biometrics for personal identification because users facial images can be captured successively from a standard video camera at a distance. Two models of the system using face recognition were tested to prevent fraud and to create automatic registration. Experimental results for a six-month test showed that the tracking capability of the method is quite practicable; the equal error rate for false rejection and false acceptance (EER) is only about one percent.