.

ISSN 2063-5346
For urgent queries please contact : +918130348310

NOVEL METHOD FOR IMPROVING ACCURACY IN DETECTING ROAD LANE WITH RECEIVER OPERATING CHARACTERISTIC USING SCALEINVARIANT FEATURE TRANSFORM OVER SUPPORT VECTOR MACHINE

Main Article Content

Rohan Raju Gorule, G. Charlyn Pushpa Latha
» doi: 10.31838/ecb/2023.12.sa1.319

Abstract

Aim: To improve the accuracy in detecting road lanes with Receiver Operating Characteristic using Novel Scale-Invariant Feature Transform over Support Vector Machine. Materials and Methods: This study contains 2 groups namely Scale-Invariant Feature Transform (SIFT) and Support Vector Machine (SVM). Each group consists of a sample size of 1506 and the study parameters include alpha value 0.05, beta value 0.2, and the power value 0.8. Their accuracies are also compared with each other using different sample sizes. Results: The Novel Scale-Invariant Feature Transform has an accuracy of 92.38% and the Support Vector Machine of 83.42% in Road Lane Detection. The significance value for performance and loss is 0.578 (p>0.05) Conclusion: The SIFT model is significantly better than the SVM in identifying Road Lane Detection. It can be also considered as a better option for the Lane Detection in General.

Article Details