Методи роботи з текстурами за допомогою засобів оброблення зображень на мові Python

Thumbnail Image
Гамаюн, Давід
Корнійчук, Максим
Journal Title
Journal ISSN
Volume Title
У статті проаналізовано наявні підходи й алгоритми оброблення зображень та подальшого їх використання для оброблення фото одягу, перетворення їх на фото текстури для 3D-моделей. Програмна частина, описана в цій статті, являє собою бекенд для онлайн-сервісу примірювання одягу на 3D-моделях у браузері, глобальним результатом якої буде тестова версія цього сервісу.
The article describes the analysis of available approaches and algorithms of image processing and their subsequent use for processing photo clothing, transforming them into photo textures for 3D models. The software part described in this article is the back-end for the online service of fitting clothes on 3D models in the browser, the global result of which will be the test version of the service. The global goal of the final product, part of which will be a practical part of this article, is to make online shopping as convenient and comfortable as possible, to minimize the need for face-to-face clothing and other types of shopping to a minimum. Every year more and more people buy a variety of things online and, of course, one of the basic human needs is clothing. But due to the fact that each person differs from each other, both in size (height, weight, waist length, etc.) and preferences, the process of buying clothes online, becomes so difficult and sometimes unpleasant. By solving the main problems that people have when buying clothes online, you can make it even more popular. There are several major problems when buying clothes online: a person does not know exactly what this or that thing will look like on him; a person is not sure whether this or that size suits him; a person wants to visually see if a certain thing will match other clothes from his wardrobe (this problem is not always solved, even by face fitting). At the initial stage, the final product will solve the last problem. To solve this problem, end users need to be able to "dress" the 3D model in the browser both in the clothes sold on the site (theoretical clothing site to which the final system will be connected) and upload their photos of clothes that will be processed into texture. for a model that will allow you to visualize how the user’s clothes will look along with the things sold on the site. After analyzing the three main technologies (OpenCV, Google Cloud Vision API, scikit-image) used for computer vision and image processing, the following conclusions have been drawn. The OpenCV library is completely free, supported by many programming languages, very well documented and has a long list of features that include exactly what you need for the task.
оброблення зображень, видалення фону, примірювання одягу в браузері, 3D у браузері, стаття, image processing, background removal, fitting clothes in the browser, 3D in the browser
Гамаюн Д. В. Методи роботи з текстурами за допомогою засобів оброблення зображень на мові Python / Гамаюн Д. В., Корнійчук М. А. // Наукові записки НаУКМА. Комп'ютерні науки. - 2020. - Т. 3. - С. 97-101.