Visualizing the Basis of Point-Base Pedestrian Detection using Pixel-wise-Attention
Pixel-wise-AttentionによるPoint型歩行者検出の判断根拠の可視化
- Delivery
- Available on this site
- Format
- Price
- Non-members (tax incl.):¥1,100 Members (tax incl.):¥880
- Publication code
- 20216146
- Paper/Info type
- Proceedings (Autumn)
No.113-21
- Pages
- 1-6(Total 6 p)
- Date of publication
- Oct 2021
- Publisher
- JSAE
- Language
- Japanese
- Event
- 2021 JSAE Annual Congress (Autumn) [Online]
Detailed Information
Author(J) | 1) 木村 秋斗, 2) 長内 淳樹, 3) 平川 翼, 4) 山下 隆義, 5) 藤吉 弘亘 |
---|---|
Author(E) | 1) Shuto Kimura, 2) Atsuki Osanai, 3) Tsubasa Hirakawa, 4) Takayoshi Yamashita, 5) Hironobu Hujiyoshi |
Affiliation(J) | 1) 中部大学, 2) 本田技研研究所, 3) 中部大学, 4) 中部大学, 5) 中部大学 |
Affiliation(E) | 1) Chubu University, 2) Honda R&D, 3) Chubu University, 4) Chubu University, 5) Chubu University |
Abstract(J) | 自動運転の安全性の向上には歩行者検出の精度向上と判断根拠の可視化が必要である.本研究は,歩行者の任意の箇所の画素を基に歩行者を検出するPoint型歩行者検出にPixel-wise-Attentionを用いることで,歩行者検出の判断根拠の可視化と精度向上を実現する. Translation |
Abstract(E) | It is necessary for the safety of automated driving technology to improve the accuracy and to visualize the basis of pedestrian detection. In our study, we use Pixel-wise-attention for Point-base object detection which detects pedestrian based on pixels at pedestrian's location. Our method improve the accuracy and visualize the basis of pedestrian detection. |