Katsuhito Fujimoto

Gesture keyboard requiring only one camera

In this paper, we propose a novel gesture-based virtual keyboard (Gesture Keyboard) of QWERTY key layout requiring only one camera. Gesture Keyboard tracks the users fingers and recognizes gestures as the input, and each virtual key of it follows a corresponding finger. Therefore, it is possible to input characters at the users preferred hand position even if displacing hands during inputting. Because Gesture Keyboard requires only one camera to obtain sensor information, keyboard-less devices can feature it easily.

High-speed, high-accuracy binarization method for recognizing text in images of low spatial resolutions

We propose a new high-speed, high-accuracy binarization method for recognizing text in document images. First character neighborhoods are extracted from input images using a global thresholding value that is shifted to the background pixel value from the thresholding value of conventional global binarization. Second, characters are extracted using an original local binarization process integrated with image interpolation. Our method takes only 1/100 the processing time over the method that performs image interpolation first. Therefore our method binarizes an A4 size text image (150dpi) in an average of only 3.3 seconds using a 166 MHz Pentium processor. Furthermore, our method reduced unrecognized characters by 46.5%, compared with conventional global binarization.

A Multi-Stage Strategy to Perspective Rectification for Mobile Phone Camera-Based Document Images

Document images captured by a mobile phone camera often have perspective distortions. Efficiency and accuracy are two important issues in designing a rectification system for such perspective documents. In this paper, we propose a new perspective rectification system based on vanishing point detection. This system achieves both the desired efficiency and accuracy using a multi-stage strategy: at the first stage, document boundaries and straight lines are used to compute vanishing points; at the second stage, text baselines and block aligns are utilized; and at the last stage, character tilt orientations are voted for the vertical vanishing point. A profit function is introduced to evaluate the reliability of detected vanishing points at each stage. If vanishing points at one stage are reliable, then rectification is ended at that stage. Otherwise, our method continues to seek more reliable vanishing points in the next stage. We have tested this method with more than 400 images including paper documents, signboards and posters. The image acceptance rate is more than 98.5% with an average speed of only about 60 ms.

Video caption duration extraction

Caption detection in the video is an active research topic in recent years. In the conventional methods, one of most difficult problems is to effectively and quickly extract the durations of the different-size captions in the complex background. To solve this problem, a novel and effective method is presented to locate and track the captions in the video. The main contributions are: (1)present a multi-scale Harris-corner based method to detect the initial position of the caption (2)propose the SGF (Steady Global Feature) to determine the caption duration. Extensive experiments demonstrate the effectiveness of the proposed method.

Degraded Character Recognition by Complementary Classifiers Combination

Character degradation is a big problem for machine printed character recognition. Two main reasons for degradation are extrinsic image degradation such as blurring and low image dimension, and intrinsic degradation caused by font variations. A recognition method that combines two complementary classifiers is proposed in this paper. The local feature based classifier extracts the local contour direction changes, which is effective for character patterns with less structure deterioration. The global feature based classifier extracts the texture distribution of the character image, which is effective when the character structure is hard to discriminate. The two complementary classifiers are combined by candidate fusion in a coarse-to-fine style. Experiments are carried on degraded Chinese character recognition. The results prove the effectiveness of our method.

A ruled-line extraction method for digital camera images

We propose a high-accuracy ruled-line extraction method for digital camera images containing shadows. The conventional method that uses adaptive binarization has a problem in that light line segments become blurred due to the adverse effect of the adaptive binarization process. Then, we propose an accurate method which uses an intentionally collapsing binary image as a clue and exploits the linearity and gray level stability of each line segment. By the experiment, we demonstrated the effectiveness of the proposed method, which reduced the extraction error by half.

Poster: Head gesture 3D interface using a head mounted camera

In this paper, we propose a real world UI that uses head gestures. This UI detects user head motion obtained in images by head mounted camera (HMC). It estimates the relative position and distance between a users head and objects user is viewing. To prevent erroneous judgment, a head-specific motion model is applied in gesture recognition. As a feedback to the user, detailed object information is displayed on head mounted display (HMD). This UI allows hands-free interaction with surrounding objects. We show the UIs effectiveness by experiments.

Logical Structure Analysis for Form Images with Arbitrary Layout by Belief Propagation

A new method for analyzing the specific logical structure of forms with unknown layout is proposed. This method uses both the target form image and a generic logical structure as inputs, and models two types of relationships probabilistically: that between strings and logical components, and that between neighboring strings having different logical components. This modeling approach allows strings to be assigned to logical components softly but robustly, and allows the use of an intuitive Bayesian probability network similar to the generic logical structure. Based on this probability network model, strings corresponding to logical components can be determined by belief propagation. This method is demonstrated to be effective by conducting tests on three types of forms.

A hybrid handwritten chinese address recognition approach

Handwritten Chinese Address Recognition describes a difficult yet important pattern recognition task. There are three difficulties in this problem: (1) Handwritten address is often of free styles and of high variations, resulting in inevitable segmentation errors. (2) The number of Chinese characters is large, leading low recognition rate for single Chinese characters. (3) Chinese address is usually irregular, i.e., different persons may write the same address in different formats. In this paper, we propose a comprehensive and hybrid approach for solving all these three difficulties. Aiming to solve (1) and (2), we adopt an enhanced holistic scheme to recognize the whole image of words (defined as a place name) instead of that of single characters. This facilitates the usage of address knowledge and avoids the difficult single character segmentation problem as well. In order to attack (3), we propose a hybrid approach that combines the word-based language model and the holistic word matching scheme. Therefore, it can deal with various irregular address. We provide theoretical justifications, outline the detailed steps, and perform a series of experiments. The experimental results on various real address demonstrate the advantages of our novel approach.

Fast precise preclassification method using a candidate table designed by projections of feature regions

For feasible recognition having many categories such as Japanese character recognition, fast matching algorithms are necessary because the matching process occupies most of recognition time. In addition, for improving recognition accuracy, the matching process must use more complicated discrimination functions or a higher dimensional feature space, which involves higher computational costs. Therefore, pre-classification is used, which outputs a set of candidate categories to decrease the number of computations of the complicated discrimination functions.