| Abstract [eng] |
In SIFT Algorithm, SVM and HOG method, after estimating the key features; we match the key features of the image and the images from the database. In case of SIFT Algorithm, we go for bounding technique where, the matched image is converted to binary image type, we then estimate the boundary of the image. In case of HOG and SVM, we extract features using HOG feature extractor and then we compared the train of data’s which are delivered to the SVM Classifier along with the feature extracted image from HOG extractor to recognize a traffic signal in the given input image. Both the Gabor filter and blob analysis uses neural network in it part of the because, we train data sets, evaluate its performance and then compare these values with other values that are kept in the data base. If the comparison matches, we select the image for the as the output image and these image will be displayed in a small message box. This comparison and recognition is done through 4 distinct algorithms and to find out which algorithm works the best to achieve the objective of road sign detection subject to speed, performance and time. |
