Our Research | 研究分野

Deep learning technique can process and optimize unknown data by learning from a lot of training data (ie, big data). For depth estimation, in general, at least two different images (ie, stereo images) are required. However, in practical applications , this requirement is not efficient. Therefore, in our research, we propose a deep learning based depth estimation from single image. ...Detail

In a conventional free-view reconstruction, the visual quality of the reconstructed 3D images is insufficient to provide 3D information to applications because of the shift and sum process. In addition, its processing speed is slow. To solve these problems, our proposed method uses a pixel rearrangement technique (PERT) with locally selective elemental images. ...Detail

In integral imaging method, captured image obtains depth information of object. By using this information, Occlusion removal technique can be performed.

Photon Counting is one of the photography method in darken situation. It obtains picture by counting number of photons. By combining this method and integral imaging, it becomes possible to shoot in place with few light sources. This feature can be applied to image acquisition in darken situation, e.f. night driving.

Imaging of transparent micro-objects is an interesting as well as a challenging problem in microscopy. Digital Holographic Microscopy system can make 3D image of human cells. We use laser to analysis cell depth information. We calculate the phase difference of two different laser beams with superior algorithm. ...Detail

3D biological imaging is 3D imaging technique for small livings observation. It performed by applying integral imaging and MALT (Moving Array-Lenslet Technique) to the microscopy. This technique can obtain 3D imformation of small cells (> 10μm) with high resolution and continuous viewpoint.