|
|
 |
| |
|
|
|
A Novel Approach to Extract Color Image Features Using Image Thinning |
|
|
|
PP: 665-672 |
|
|
|
doi:10.18576/amis/160501
|
|
|
|
Author(s) |
|
|
|
M. Abu-Faraj,
Z. A. Alqadi,
B. Al-Ahmad,
K. Aldebei,
B. J. A. Ali,
|
|
|
|
Abstract |
|
|
| Color digital image recognition systems require discrimination processes that have high efficiency and speed to make the appropriate decision. Since the size of most digital color images is large, which may negatively affect the efficiency of the discrimination system, it is necessary to represent the digital image with a small set of values called the features vector. These feature vectors can be used instead of the image to identify the person using his image. In this research paper, we present a simplified and easy-to-implement method for extracting the features vector of any colored digital image. We use some morphological operations to thin the digital image and then calculate the thinning ratio as a feature of the colored digital image. The results of this paper show that the number of feature vectors depends on the number of structuring elements used in the digital image thinning operations. These structuring elements are replaced to form new image features. For efficiency comparisons, the results are compared with the K-means clustering method to show the speedup provided by the proposed method.
|
|
|
|
|
|
|
|
