ANALYSIS AND COMPARISON OF METHODS FOR IMPROVING CONTRAST OF IMAGES
A.M. Varenko · Scientific notes of Taurida National V I Vernadsky University Series Technical Sciences · 2019
Since the end of the last century, for a number of reasons, digital video cameras have begun to gain widespread popularity.For the sake of increasing demand for video cameras, they reduced the cost price, which significantly affected the quality of video and photos.It is possible to slightly improve the quality using various types of filters: high-frequency, low-frequency, median, etc. and combining them.This allows to highlight the main objects and details in the image and pick up noise or extra details.Due to this, it is possible to improve the quality of video and photos obtained using a video camera, when it is not possible to use the best optical matrix or add brightness using additional light sources.Currently, there is a widespread problem of improving the visual quality of the image for various purposes: finding a moving object, reducing noise, highlighting the object against the background and for other purposes.Also, increasing the contrast of the image can be used by attackers to encrypt secret content and transfer it.Changes to the image can rarely be detected without additional work with the image.But most of the changes are clearly visible on the histogram of the image, which has undergone changes.Despite the fact that a number of algorithms have been proposed for detecting the content encoded in the image, the reliability of the methods when using treatments of various types of images is unsatisfactory.This article provides an overview of literary sources, the authors of which offer two options for obtaining the original image for further revealing hidden content or simply obtaining the original source image using convolutional neural networks of various types, gamma correction, and a high-pass filter; the analysis of the effectiveness of each of the methods and the scope of application of these methods is done.Using the software developed in the Java programming language, an arbitrary image is analyzed from its own image library, histograms are constructed for each of the component image colors, a color image is processed using a highpass filter; image processing by convolutional artificial neural networks of two types is carried out.