Graduate School of
Science and Engineering

Our major research efforts have been focused on computer vision and human-computer interaction for medical images. The goal of our researches is to let computers help physicians in the image interpretation process. In computer vision, we have been developing image processing techniques such as a filter bank for enhancing lesions in medical images, a subtraction technique for visualizing lesions, and a super-resolution technique for improving image resolution and image quality. In human-computer interaction, we have been developing a computerized scheme for evaluating the likelihood of malignancy and/or histological classifications on lesion by using pattern recognition and artificial intelligence. We also investigate if the proposed methods are useful for improving the diagnostic accuracy and for reducing the interpretation time in observer study and clinical practice.