Improving Animal Recognition Accuracy using Image De-fencing

Kohei Kawanaka, Rabarison M. Felicia, Hiroki Tanioka, Masahiko Sano, Kenji Matsuura, Tetsushi Ueta

Abstract: When a computer recognizes an animal on an image, this paper, we aim to improve the recognition rate of objects an accuracy of animal recognition decreases if a part of the animal's body is covered by fences. On the other hand, when a human recognizes an object on the image, it is known to be possible to remove the fences and supplement the removed areas, and recognize the object, even if a part of the object is covered by fences. In this research, we aim to improve the accuracy of hidden by fences. animal recognition using de-fencing on computer.