Name
Abstract | ||
|---|---|---|
Given the tremendous attention on autonomous driving technology, traffic sign detection is increasingly important to guide the driving of the vehicles. However, the existed object detection methods cannot be directly employed as they always ignore the small objects. Different from general targets, the small objects only occupy a few pixels in images, which makes it hard to extract the discriminative features from them. In this paper, we propose a LOcal COntext based Faster R-CNN (LOCO) approach for traffic sign detection, which utilizes the regional proposal network for proposal generation, and local context information surrounding proposals for classifying. More importantly, a local context layer is designed to automatically extract the discriminative information from the regions around the proposal objects. The evaluations on two public real-world datasets demonstrate that our approach can significantly outperform the state-of-the-art methods. |
Year | DOI | Venue |
|---|---|---|
2018 | 10.1007/978-3-319-73603-7_27 | Lecture Notes in Computer Science |
Keywords | Field | DocType |
Traffic sign detection,Small object,Local context,Faster R-CNN,Autonomous driving | Object detection,Computer vision,Pattern recognition,Computer science,Context based,Artificial intelligence,Pixel,Discriminative model,Traffic sign detection | Conference |
Volume | ISSN | Citations |
10704 | 0302-9743 | 1 |
PageRank | References | Authors |
0.35 | 24 | 4 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Peng Cheng | 1 | 7 | 1.11 |
| Wu Liu | 2 | 275 | 34.53 |
| Yifan Zhang | 3 | 2 | 1.72 |
| Huadong Ma | 4 | 2020 | 179.93 |