書誌事項
- タイトル別名
-
- Motion Estimation of Visual Target Using Event-Based Vision and Spiking Neural Network
抄録
<p>We proposed a method for estimating the angular velocity of a rotating visual target robustly to changes in illumination conditions using event-based vision and spiking neural networks (SNNs). By using the output of the event-based vision directly as input to the SNN, the image processing function can be achieved by keeping the advantages of the vision sensor in the previous stage. A board printed with natural images was placed at a certain distance from the vision system as a viewing target. The vision system captured rotating objects at various angular velocities under multiple lighting conditions to generate a data set. Using this data set, we verified whether the SNN can estimate the visual target. We achieved about 80% accuracy on both the training and test datasets in both bright and dark lighting environments.</p>
収録刊行物
-
- ロボティクス・メカトロニクス講演会講演概要集
-
ロボティクス・メカトロニクス講演会講演概要集 2021 (0), 2P2-H05-, 2021
一般社団法人 日本機械学会
- Tweet
詳細情報 詳細情報について
-
- CRID
- 1390572012411521024
-
- NII論文ID
- 130008135692
-
- ISSN
- 24243124
-
- HANDLE
- 10228/00009043
-
- 本文言語コード
- ja
-
- データソース種別
-
- JaLC
- IRDB
- Crossref
- CiNii Articles
-
- 抄録ライセンスフラグ
- 使用不可