[in Japanese]

View full abstract

It is difficult for gaze detection systems to detect pupils in camera images containing eyeglasses reflections caused by the infrared light sources necessary for gaze detection. In a previous study with a system using two cameras, we proposed a geometrical pupil detection algorithm. However, depending on the positional relationship between the pupil and the reflection, the algorithm was not always effective. In the present study, we arranged three cameras—one at each vertex of a triangle—and composed three camera pairs with different base line directions. We propose several new algorisms, and in the basic algorithm, pupil candidates indicating the suprathreshold angle between two planes, each of which includes the cameras composing a camera pair and the candidate detected by one of the two cameras, are removed from the candidate list. A significant improvement of pupil detection is experimentally demonstrated.

View full abstract

As 3D display devices are becoming more popular, automatic editing methods such as inpainting for stereoscopic images have grown increasingly important. We propose an inpainting method for stereoscopic images by minimizing an energy function. We expand the conventional image inpainting method by adding a constraint condition that the corresponding pixel has the same intensity, thus enabling the proposed method to deliver more natural result. The proposed method can also be applied if there is an object with different visibility between the right and left input images by using the depth value of the other for depth map inpainting. Experimental results demonstrated the efficiency of our method.

View full abstract

[in Japanese]

View full abstract

Identifying subtle human facial expressions is important in human interactions such as human-robot communications. We propose a method to classify two types of facial expression, neutral (expressionless) and the subtle expression of happiness, using an image acquired with a camera. In the proposed method, facial shape features such as wrinkle-angle or unevenness of face surface are extracted via Gabor filters in multiple ROIs on a human face. High angle resolution Gabor filters are used to detect subtle facial expressions and the positions of the detection window and Gabor filter parameters are optimized by a learning process using AdaBoost. Changes to facial expression are estimated by comparing an input image to a learned database with the features. Experimental results showed that the proposed system has a 0.84 precision rate and a 0.91 recall rate in the case of detecting subtle changes.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract