TY  - JOUR
T1  - An Intelligent Image Classifier Based on Histogram of Oriented Gradients Features
AU - Talib Gaata, Methaq AU - Muneam, Sahar 
JO  - Journal of Engineering and Applied Sciences
VL  - 13
IS  - 20
SP  - 8506
EP  - 8510
PY  - 2018
DA  - 2001/08/19
SN  - 1816-949x
DO  - jeasci.2018.8506.8510
UR  - https://makhillpublications.co/view-article.php?doi=jeasci.2018.8506.8510
KW  - Image classification
KW  -feature extraction
KW  -ANN
KW  -(HOG)
KW  -extract vector
KW  -high accuracy
AB  - This study presents an intelligent classifier for images classification based on Artificial Neural
Network (ANN). The Histogram of Oriented Gradients (HOG) techinque has been used in order to extract
features from image. The ANN supervised feed-forward scaled conjugate gradient algorithm used to bulid the
proposed classifier. The input image is processed directly to extract vector features regardless of size or colour
map. The architecture of ANN is selected to be simple and appropriate to carry out the classification process
with high accuracy. This work is performed on the Caltech dataset. Four classes of image are used to test and
evaluate the performance of the proposed classifier (96 images for all category), the testing images consists of
192 images (48 images for each category). Experimental results showed that the classification rate was 93.23%.
ER  -