| Presentation | 2019-12-20 An Efficient Block-wise Object Detection Method using Consecutive Frames for High Resolution Video Kazuki Hozumi, Yoichi Tomioka, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In recent years, in the fields such as surveillance cameras and in-vehicle camera systems, efficient deep-learning-based object detection methods, such as Single Shot MultiBox Detector (SSD), that do not require window scanning have received a significant attention. However, these methods require a lot of memory and computation. For this reason, when we apply them to higher definition video, it can be necessary to divide the video into multiple blocks for inference processing due to restrictions on memory capacity of GPUs or FPGAs. However, the detection accuracy of objects near the block division boundary can be low. Although we can use overlap blocks to reduce the effects of block boundary, it increases the number of blocks and execution time. In this paper, we propose a method for reducing the execution time per frame, which assigns a different pattern to each fame and integrates the results of object detection from multiple frames. In the experiments, the object detection accuracy was evaluated using three data from the Multiple Object Tracking Benchmark dataset 2017. We reduced the number of blocks per frames to 55.6% while the accuracy denegeration is within 4.5%. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Convolutional neural network / Deep learning / Object detection / Single Shot Multibox Detector Detector |
| Paper # | PRMU2019-57 |
| Date of Issue | 2019-12-12 (PRMU) |
| Conference Information | |
| Committee | PRMU |
|---|---|
| Conference Date | 2019/12/19(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Yoichi Sato(Univ. of Tokyo) |
| Vice Chair | Toru Tamaki(Hiroshima Univ.) / Akisato Kimura(NTT) |
| Secretary | Toru Tamaki(NTT) / Akisato Kimura(OMRON SINICX) |
| Assistant | Yusuke Uchida(DeNA) / Takayoshi Yamashita(Chubu Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Pattern Recognition and Media Understanding |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | An Efficient Block-wise Object Detection Method using Consecutive Frames for High Resolution Video |
| Sub Title (in English) | |
| Keyword(1) | Convolutional neural network |
| Keyword(2) | Deep learning |
| Keyword(3) | Object detection |
| Keyword(4) | Single Shot Multibox Detector Detector |
| 1st Author's Name | Kazuki Hozumi |
| 1st Author's Affiliation | University of Aizu(UoA) |
| 2nd Author's Name | Yoichi Tomioka |
| 2nd Author's Affiliation | University of Aizu(UoA) |
| Date | 2019-12-20 |
| Paper # | PRMU2019-57 |
| Volume (vol) | vol.119 |
| Number (no) | PRMU-347 |
| Page | pp.pp.69-74(PRMU), |
| #Pages | 6 |
| Date of Issue | 2019-12-12 (PRMU) |