Optimization of Rotation-Invariant Feature Detection Method for Pedestrian Recognition

Open Access

Abstract: In the technology of Advanced Safety Vehicle (ASV), it is an important element to detect
pedestrians by vision. However, pedestrians have pose variations such as a rotation. That’s why we cannot always get similar feature values. In order to solve this problem, we use the rotation-invariant histogram of oriented gradient (RI-HOG), but this method needs very high costs because it’s not optimized for pedestrian recognition. Therefore, we improved calculation method of RI-HOG for pedestrian recognition and compared this proposed method with the conventional method. As a result, recognition rate was increased by about 20%.

Keywords: Pedestrian recognition; rotation-invariant descriptor; optimal basis functions

Toshiki Yahiro and Kousuke Matsushima

The Author field can not be Empty

NA

The Institution field can't be Empty

Volume 1 Issue 1

Volume and Issue can't be empty

38 - 43

The Page Numbers field can't be Empty

25-12-2015

Publication Date field can't be Empty

Description

Abstract: In the technology of Advanced Safety Vehicle (ASV), it is an important element to detect
pedestrians by vision. However, pedestrians have pose variations such as a rotation. That’s why we cannot always get similar feature values. In order to solve this problem, we use the rotation-invariant histogram of oriented gradient (RI-HOG), but this method needs very high costs because it’s not optimized for pedestrian recognition. Therefore, we improved calculation method of RI-HOG for pedestrian recognition and compared this proposed method with the conventional method. As a result, recognition rate was increased by about 20%.

Keywords: Pedestrian recognition; rotation-invariant descriptor; optimal basis functions