Research on Structured Feature Extraction and Matching Method for Video Retrieval on Non-Rigid Objects
| Project/Area Number |
15K00251
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Perceptual information processing
|
|---|
| Research Institution | Aichi Institute of Technology |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2015-04-01 – 2018-03-31
|
| Project Status |
Completed (Fiscal Year 2017)
|
| Budget Amount *help |
¥4,550,000 (Direct Cost: ¥3,500,000、Indirect Cost: ¥1,050,000)
Fiscal Year 2017: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2016: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2015: ¥2,080,000 (Direct Cost: ¥1,600,000、Indirect Cost: ¥480,000)
|
|---|
| Keywords | 映像検索 / 非剛体 / 変形 / 特徴抽出 / 特徴照合 |
|---|
| Outline of Final Research Achievements |
This research has engaged on the technology of video retrieval for non-rigid objects with flexible shape such as cloths and flags. National flags has been used as non-rigid object class in this research.
A proposed detection method in the research selects frame images suitable for retrieval from the video with deforming flags during the scene and detects the region of flags in selected images. Deformable Part Model (DPM), one of machine learning technology, is used for the selection and detection by learning characteristics of national flags with various shape deformation, and experiments showed the effectiveness of the proposed detection method.
Another proposed recognition method in the research devides the deteced flag region into subregions, extracts the color mean vectors as features for all subregions, matches them with features in DB and computes the rank result. Experiments also showed the effectiveness of the proposed recognition method.
|
Report
(4 results)
Research Products
(4 results)
