A Hybrid Approach for Efficient Traffic Sign Detection Using Yolov8 And SAM
Open Access
- 22 March 2024
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 298-302
Abstract
In this study, we present an innovative hybrid approach for traffic sign detection in autonomous driving, combining YOLOv8’s real-time detection capabilities with the Segment Anything Model (SAM), enhanced through Visual Prompt Tuning. This methodology addresses the challenge of accurately identifying diverse traffic signs, particularly those that are less common or visually complex. After the initial detection by YOLOv8, the SAM (Segment Anything Model), enhanced with Visual Prompt Tuning, is then applied to refine and improve the accuracy of these detections, especially for signs that are less common or have complex visual features. This two-step process, combining the quick detection capabilities of YOLOv8 with the detailed segmentation ability of SAM, aims to achieve high accuracy in traffic sign recognition, which is essential for the safe operation of autonomous vehicles.Keywords
Funding Information
- Guangzhou Science and Technology Plan Project (2024B01W0079, 202206030005, 202206010056)
- Nansha Key RD Program (2022ZD014)
- Science and Technology Planning Project of Guangdong Province (2023B1212060029)
- Guangdong Basic and applied Research Foundation (2021B1515120032, 2022B1515120072)
This publication has 11 references indexed in Scilit:
- Segment Anything2023
- SegGPT: Towards Segmenting Everything In Context2023
- YOLOv7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors2023
- A real-time and high-precision method for small traffic-signs recognitionNeural Computing and Applications, 2021
- Feature Refinement and Filter Network for Person Re-IdentificationIEEE Transactions on Circuits and Systems for Video Technology, 2020
- SADANet: Integrating Scale-Aware and Domain Adaptive for Traffic Sign DetectionIEEE Access, 2020
- Improved Faster R-CNN Traffic Sign Detection Based on a Second Region of Interest and Highly Possible Regions Proposal NetworkSensors, 2019
- A Real-Time Chinese Traffic Sign Detection Algorithm Based on Modified YOLOv2Algorithms, 2017
- Traffic-Sign Detection and Classification in the Wild2016
- DeepID-Net: Deformable deep convolutional neural networks for object detection2015