Development of an Algorithm for Extracting Features in Images Based on the Gabor Filter
Abstract
This article discusses various filters such as Laplace, Median, Gaussian, Sobel, and Canny Edge detectors that are still widely used today for image feature extraction. These filters serve quite important functions in areas such as computer vision and medical diagnostics by detecting edges, shapes, and other significant features of an object. Each filter has specific characteristics: Laplace detects edges using second derivatives; Median filters reduce noise, especially impulse noise; Gaussian filters smooth images to reduce high-frequency noise; and Sobel filters help detect orientation-based edges using gradient derivatives. In addition, the Canny Edge detector stands out for its edge detection accuracy using noise reduction and gradient analysis. The article also discusses the development of a new algorithm based on the mathematical model of the Gabor filter, which offers improved edge detection accuracy in images. The Gabor filter, known for its multi-directional and multi-dimensional flexibility, allows for detailed frequency and direction analysis, proving its high efficiency in complex feature extraction. The study demonstrates the superiority of the Gabor filter in the accuracy and efficiency of feature extraction, especially in applications that require accurate edge detection in various types of images. The proposed approach highlights the usefulness of the Gabor filter in achieving highquality results in image analysis tasks.