Scale Invariant Mask R-CNN for Pedestrian Detection
Pedestrian detection is a challenging and active research area in computer vision. Recognizing pedestrianshelps in various utility applications such as event detection in overcrowded areas, gender, and gaitclassification, etc. In this domain, the most recent research is based on instance segmentatio...
| Authors: | , , |
|---|---|
| Format: | article |
| Publication Date: | 2020 |
| Country: | España |
| Institution: | Universitat Autònoma de Barcelona |
| Repository: | Dipòsit Digital de Documents de la UAB |
| Language: | English |
| OAI Identifier: | oai:ddd.uab.cat:233664 |
| Online Access: | https://ddd.uab.cat/record/233664 https://dx.doi.org/urn:doi:10.5565/rev/elcvia.1278 |
| Access Level: | Open access |
| Keyword: | Convolutional neural network Instance segmentation Pedestrian Detection Mask R-CNN |
| Summary: | Pedestrian detection is a challenging and active research area in computer vision. Recognizing pedestrianshelps in various utility applications such as event detection in overcrowded areas, gender, and gaitclassification, etc. In this domain, the most recent research is based on instance segmentation using MaskR-CNN. Most of the pedestrian detection method uses a feature of different body portions for identifying aperson. This feature-based approach is not efficient enough to differentiate pedestrians in real-time, wherethe background changing. In this paper, a combined approach of scale-invariant feature map generationfor detecting a small pedestrian and Mask R-CNN has been proposed for multiple pedestrian detection toovercome this drawback. The new database was created by recording the behavior of the student at theprominent places of the engineering institute. This database is comparatively new for pedestrian detectionin the academic environment. The proposed Scale-invariant Mask R-CNN has been tested on the newlycreated database and has been compared with the Caltech [1], INRIA [2], MS COCO [3], ETH [4], andKITTI [5] database. The experimental result shows significant performance improvement in pedestrian detection as compared to the existing approaches of pedestrian detection and instance segmentation. Finally, we conclude and investigate the directions for future research. |
|---|