Simple, Flexible Method to Extract Digital Image Features

Abstract

The process of digital image features extraction is very important and it is required in many applications such as classification, prediction and regression. The extracted features for each image must be unique and capable to be used as an image identifier. In this paper we will introduce a method of image features extraction; it will be shown that this method will enhance the efficiency of the features extraction process. The proposed method will be experimentally tested using various images; the obtained experimental results will be compared with other existing methods of feature extraction to show the advantages of the proposed method and to show how to increase the speed up of the method.

Enhanced Average Filtering Technique for Mitigating Salt and Pepper Noise in High-Resolution Color Images

“…The results of the proposed method were compared with the results of the filters in studies [1][2][3][4][5][6], and the proposed method gave a significant enhancement of PSNR for any image and any NR value.…”

“…Digital images, comprising a substantial complex [1] of data organized into a three-dimensional matrix, as depicted in Figure 1, allocate the first binary matrix to represent the red color, the second for the green color, and the third for the blue color [2][3][4].…”

Autoregressive prediction analysis using machine deep learning

“…The regression means square error (MSE) between the calculated and the target outputs may be high, causing a poor effective prediction. thus, the machine deep learning approach can be used to minimize MSE and enhance the prediction process accuracy [6], [7]. The machine deep learning (MDL) approach is a process of using an artificial neural network (ANN) [8]- [10] as a prediction tool.…”