Ryosuke Kasahara

Decision Feedback Partial Response Maximum Likelihood for Super-Resolution Media

A decision feedback partial response maximum likelihood (PRML) for super-resolution media was developed. Decision feedback is used to compensate for nonlinear distortion in the readout signals of super-resolution media, making it possible to compensate for long-bit nonlinear distortion in small circuits. An field programmable gate array (FPGA) was fabricated with a decision feedback PRML, and a real-time bit error rate (bER) measuring system was developed. As a result, a bER of 4×10-5 was achieved with an actual readout signal at the double density of a Blu-ray disc converted to the optical properties of the experimental setup using a red-laser system. Also, a bER of 1.5×10-5 was achieved at double the density of an a high definition digital versatile disc read-only memory (HD DVD-ROM), and the radial and tangential tilt margins were measured in a blue-laser system.

Noninvasive glucose monitoring using mid-infrared absorption spectroscopy based on a few wavenumbers

A method for performing noninvasive blood glucose measurements was developed. The method is based on mid-infrared absorption spectroscopy and uses only a few wavenumbers to measure blood glucose levels in vivo unconditionally. We found that the regression of blood glucose levels using only three wavenumbers, which were selected using a series cross-validation technique, realized accuracies comparable to those of cases in which a greater number of wavenumbers are used. In addition, we demonstrated the performance of this model through correlations among different types of data.

Improvement of image quality by polarization mixing

Information about the polarization of light is valuable because it contains information about the light source illuminating an object, the illumination angle, and the object material. However, polarization information strongly depends on the direction of the light source, and it is difficult to use a polarization image with various recognition algorithms outdoors because the angle of the sun varies. We propose an image enhancement method for utilizing polarization information in many such situations where the light source is not fixed. We take two approaches to overcome this problem. First, we compute an image that is the combination of a polarization image and the corresponding brightness image. Because of the angle of the light source, the polarization contains no information about some scenes. Therefore, it is difficult to use only polarization information in any scene for applications such as object detection. However, if we use a combination of a polarization image and a brightness image, the brightness image can complement the lack of scene information. The second approach is finding features that depend less on the direction of the light source. We propose a method for extracting scene features based on a calculation of the reflection model including polarization effects. A polarization camera that has micro-polarizers on each pixel of the image sensor was built and used for capturing images. We discuss examples that demonstrate the improved visibility of objects by applying our proposed method to, e.g., the visibility of lane markers on wet roads.