Automatic face recognition in real-time
Imagine segmentation recognition using convolution neural networks (CNN) is increasing its popularity especially after the IMAGENET. Despite the state-of-the-art performance, CNN demands huge computational load that limits its applications in real-time environments. To address these issues, recently...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
2019
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/78345 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Imagine segmentation recognition using convolution neural networks (CNN) is increasing its popularity especially after the IMAGENET. Despite the state-of-the-art performance, CNN demands huge computational load that limits its applications in real-time environments. To address these issues, recently, You Only Look Once (YOLO), a variant of CNN, is developed to achieve the comparable performance of CNN with significantly less computational resources. In this project, we aim to employ the YOLO architecture as the core component to develop the real-time identification of number of persons in a given room at every time instant. Furthermore, the proposed architecture uses the image segmentation information obtained from YOLO to tag the persons in real-time. To validate the proposed architecture in real-world settings, YOLO along with the tagging algorithm is implemented in NVIDIA Jetson-TXII board. Results showed that the proposed architecture can successfully recognize the number of persons in a given along with their name-tags. |
|---|