Performance analysis on road sign detection, extraction and recognition techniques
| Autor: | E. Shoba, A. Suruliandi |
|---|---|
| Rok vydání: | 2013 |
| Předmět: |
Contextual image classification
Computer science business.industry Feature extraction ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Cognitive neuroscience of visual object recognition Pattern recognition Image segmentation Object detection Histogram Computer vision Segmentation Artificial intelligence business Sign (mathematics) |
| Zdroj: | 2013 International Conference on Circuits, Power and Computing Technologies (ICCPCT). |
| Popis: | Automatic detection and recognition of road traffic signs is an important work for regulating the traffic and guiding and warning drivers and pedestrians. The main aim of this work is to compare the performance of the road sign detection, extraction and recognition techniques and find the best one. This project investigates two techniques. In the first approach, Color Classification is used for detecting the road sign. After extraction, Wavelet based classification is used for recognize the road sign. In the second approach, Color Segmentation and Shape Classification is used for detect the road sign. After detecting the road sign Local energy based shape histogram (LESH) is used for recognize the road sign. To evaluate the performance of the above two approaches Recognition Rate (RR) metric is used. From the evaluation, we conclude that the performance of LESH method provides better performance from the overall results. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|