Highly accurate pedestrian recognition for automotive applications
| Project/Area Number |
20700048
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
Computer system/Network
|
|---|
| Research Institution | Nara Institute of Science and Technology |
|---|
Principal Investigator |
MIYAMOTO Ryusuke Nara Institute of Science and Technology, 情報科学研究科, 助教 (10452525)
|
|---|
| Project Period (FY) |
2008 – 2010
|
| Project Status |
Completed (Fiscal Year 2010)
|
| Budget Amount *help |
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2010: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000)
Fiscal Year 2009: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2008: ¥2,080,000 (Direct Cost: ¥1,600,000、Indirect Cost: ¥480,000)
|
|---|
| Keywords | 画像認識 / 組込みシステム / 車載 / 歩行者認識 / 歩行者検出 / 歩行者追跡 |
|---|
| Research Abstract |
In this research, we propose the parallel processing strategy suitable for NVIDIA GPUs, one of the most widely used multi-core processors, and FPGA to achieve real-time processing of the CoHOG-based pedestrian detection, which is required for accurate recognition in automotive applications. In addition, a novel computing scheme using MCMC which can reduce computational cost without degradation of detection accuracy. For a tracking scheme, which is indispensable for accurate recognition, a multiple cue-based scheme is proposed and its parallel implementation is also shown. Finally, a recognition scheme combining the detection and the tracking schemes is proposed, and it is shown that the highly accurate recognition scheme can be computed in real time.
|
Report
(4 results)
Research Products
(64 results)
