Performance Analysis of Symbolic Road Recognition for On-Road Driving
Mike Foedisch, Craig I. Schlenoff, Rajmohan Madhavan
Previous approaches to road sensing, namely road detection were based on segmenting the sensor data, i.e. color camera image, into road and non-road areas. Performance evaluation for such algorithms could be performed in a relatively straightforward fashion by comparing the algorithm s result with ground truth. Ground truth for such an image-based evaluation approach could be limited to a geometrical structure describing the road area in the original image. However, the development of our new highlevel road sensing approach, which is a model-based approach to road recognition, makes new demands to performance analysis and subsequent performance evaluation which would include comparison with ground truth. In this paper1, we will briefly describe the new road recognition approach, show performance analysis results and discuss performance evaluation issues.
Performance Metrics for Intelligent Systems Workshop| 2006 | PerMIS
August 21-23, 2006
Performance Metrics for Intelligent Systems (PerMIS) Workshop 2006
, Schlenoff, C.
and Madhavan, R.
Performance Analysis of Symbolic Road Recognition for On-Road Driving, Performance Metrics for Intelligent Systems Workshop| 2006 | PerMIS, Gaithersburg, MD, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=823594
(Accessed June 2, 2023)