Small object recognition using machine learning techniques
Self-driving cars developments are becoming the norm in both automotive and IT industry. Recent researches were focusing on ways these autonomous vehicles interact with its environment. Object detection using cameras is one of the important vision aids and they are used for applications such as path...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
2018
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/75496 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Self-driving cars developments are becoming the norm in both automotive and IT industry. Recent researches were focusing on ways these autonomous vehicles interact with its environment. Object detection using cameras is one of the important vision aids and they are used for applications such as path planning, object avoidance and localizations. In this study, the particular case of You Only Look Once v2 on the Udacity crowdAI data set was analyzed for its performance on different sizes of object classes. This work improves on the object performance of the original detection model for Car by an average of 12% on the VOC2007 Average Precision for all sizes. Specifically, the Small-Car and Small-Truck improved by 2.8% and 4% on the same metric. |
|---|