Subject Area
Electrical Engineering
Article Type
Original Study
Abstract
Shape recognition is an essential component in any computer vision system. A general technique based on the normalized area vector concept is presented for solving the problems of distortion and invariance against translation, rotation and zoom. À new algorithm is described for extraction of contours of complex nonoverlapping objects which may contain any number of holes of any shape and in any configuration. Experimental results show the robustness of the proposed technique in comparison with: other well known techniques such as the shape Dumber. A training set of twenty patterns is used for generation of reference patterns. Á test set of tenty four patterns is used to show the performance of the proposed technique and resulted in a hundred percentage of Correct classification for both the training and test set patterns.
Recommended Citation
Hamed, M.; Tolba, Ahmed; and El-Hendy, H.
(2021)
"A General Technique for Recognition of Planar Shapes.,"
Mansoura Engineering Journal: Vol. 18
:
Iss.
3
, Article 8.
Available at:
https://doi.org/10.21608/bfemu.2021.165527
