2021 Fiscal Year Final Research Report
Hierarchical Person Identification using Individual and Group Features
| Project/Area Number |
18K11361
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 61010:Perceptual information processing-related
|
|---|
| Research Institution | Tsuda University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
浜本 隆之 東京理科大学, 工学部電気工学科, 教授 (10297624)
|
|---|
| Project Period (FY) |
2018-04-01 – 2022-03-31
|
| Keywords | 知覚情報処理 / 画像認識 |
|---|
| Outline of Final Research Achievements |
We develop a method for person identification from surveillance videos. Conventional methods used color and shape features of individual person. However, the information that can be extracted from individual tends to be limited; thus, there is a certain limit to the improvement of identification performance. In addition, when the number of people to be identified increases, the identification performance generally deteriorates. In this research, we investigate a hierarchical identification approach for groups and individuals. If we can analyze the information of a group, we can use the features related to the group. Furthermore, if the group can be identified, the number of candidates to be identified can be decreased because it can be considered that groups contain only a few people. Based on this hierarchical identification approach, we develop techniques of a person re-identification and person identification.
|
| Free Research Field |
画像認識
|
| Academic Significance and Societal Importance of the Research Achievements |
映像からの人物照合技術の高度化は，安心・安全な社会環境の実現において必要不可欠である．人物個人だけでなく集団の情報を照合に活用する観点は，これまでの従来技術にないものである．また，集団・個人の階層的認識は，照合対象の増大に伴う性能低下という当該分野の本質的な課題の解決を試みる独創的なアプローチである．このように，本研究課題で検討および開発した技術は，当該分野にインパクトを与えるものであると考える．
|
