| Authors |
M. Hori, M. Kinoshita, T. Yamazato, H. Okada, T. Fujii, K. Kamakura, T. Yendo, S. Arai |
| Title |
An LED Transmitter Detection using Linear SVM and CNN for ITS Image Sensor Communication |
| Authority |
3rd International Conference and Exhibition on Visible Light Communications (ICELVC) |
| Summary |
Image sensor communication (ISC) is a type of visible light communications (VLC) that has high affinity with the field of intelligent transport systems (ITSs). In an ISC, accurate detection of the transmitter from a captured image is critical, because the receiver uses pixels that sense VLC signal for data reception. The purpose of this study is to reduce the false-positive and the false-negative probabilities of the transmitter. To achieve this goal, we propose a novel LED transmitter detection method composed of two stages: the candidate extraction stage by a linear support vector machine (SVM) and the classification stage by a convolutional neural network (CNN). Then, we show that the proposed method is robust to vehicle vibration and other noises such as non-transmitter LED compared to the conventional method. |
| 年月 |
2019年3月 |
| DOI/Handle |
|
| 開催場所 |
Seoul, Korea |
| 研究テーマ |
可視光通信／光無線通信
高度交通システム（ＩＴＳ）
|
| 言語 |
英語 |
| 原稿/プレゼン資料 |
 /  (ローカル限定) |
