|Affiliation||Department of Computer Science and Engineering|
|Concurrent post||Research Center for Future Vehicle City|
|Fields of Research||Computer Vision / Image Processing|
|Degree||Dr. Eng. (Osaka University)|
|Academic Societies||IEICE / IPSJ / IEEE|
Please append ".tut.ac.jp" to the end of the address above.
|Laboratory website URL||http://www.img.cs.tut.ac.jp/~kanazawa/|
|Researcher information URL（researchmap）||Researcher information|
Computer vision is a very important tool for obtaining various informations of a scene, such as 3-D shapes, 2-D shapes, moving information of objects, colors of objects and lightes, and so on. In particular 3-D shape information is very useful for us. However, the accuracies of the reconstructed 3-D shapes are often very low when we compute them using real scene data. So, we study for accurate methods for matching images and reconstructing 3-D shapes with real images. We also study for computer vision applications to the real world.
Image matching is an important process for many computer vision methods. We focus on the cases: wide baseline, images without texture, and images including repetitive patterns.
Title of class
Linear Algebra 2 / Machine Learning and Pattern Recognition / Advanced Image Processing