A Novel Shape Descriptor for Object Recognition
Keywords:
Shape Descriptor, nesne tanıma, MNISTAbstract
In this study a novel shape descriptor for object recognition is proposed. As a preprocessing stage, Canny edge detection [4] is applied to input images. Output of Canny edge detector, namely edge image, is sampled and various number of points are selected. Chosen points are input to the new shape descriptor. Proposed shape descriptor is composed of deviations from average range and average angle. Shape descriptor is used as a feature extractor output of which is fed to linear classifier. Linear classifier is trained using pseudo-inverse and gradient descent techniques. Full MNIST dataset is used to test the system and results are reported.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2023 International Journal of Computational and Experimental Science and Engineering
![Creative Commons License](http://i.creativecommons.org/l/by/4.0/88x31.png)
This work is licensed under a Creative Commons Attribution 4.0 International License.