Multi-Attention-Guided Cascading Network for End-to-End Person Search
The key procedure is to accurately identify pedestrians in complex scenes and effectively embed features from multiple vision cues. However, it is still a limitation to coordinate two tasks in the unified framework, thus leading to high computational overhead and unsatisfactory search performance. F...
保存先:
| 主要な著者: | , |
|---|---|
| フォーマット: | Artigo |
| 言語: | Inglês |
| 出版事項: |
MDPI AG
2023-04-01
|
| シリーズ: | Applied Sciences |
| 主題: | |
| オンライン･アクセス: | https://www.mdpi.com/2076-3417/13/9/5576 |
| タグ: |
タグなし, このレコードへの初めてのタグを付けませんか!
|