Research on Image Recognition Using The Viewpoint of Test Sample
| Project/Area Number |
15K00252
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Perceptual information processing
|
|---|
| Research Institution | Meijo University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2015-04-01 – 2018-03-31
|
| Project Status |
Completed (Fiscal Year 2017)
|
| Budget Amount *help |
¥4,680,000 (Direct Cost: ¥3,600,000、Indirect Cost: ¥1,080,000)
Fiscal Year 2017: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Fiscal Year 2016: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2015: ¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000)
|
|---|
| Keywords | ディープラーニング / 適応的 / 画像認識 / 対象計数 / 対象識別 / 対象追跡 / セマンティックセグメンテーション / 画像ラベリング |
|---|
| Outline of Final Research Achievements |
We proposed adaptive integration method of multiple CNNs. When there is large appearance changes, the accuracy tend to drop. We used gating CNN and expert CNNs. Gating CNN is for assigning sub-task to expert CNNs. Expert CNN trained images assigned by the gating CNN. The effectiveness of the proposed method is shown by experiments.
In the task on object tracking in a video, the tracking accuracy drops when occlusion occurs. We proposed a method to predict whether occlusion occurs or not. When occlusion occurs, the learning rate of tracking model is changed. Our method can track object under the occlusion with higher accuracy in comparison with conventional method.
|
Report
(4 results)
Research Products
(41 results)
