A Color Channel Based Analysis on Image Tessellation

Abstract

In this thesis, it is investigated whether separate triangulation of the RGB components that make up the image would be more efficient in terms of size and quality than direct triangulation of the main image. Different tesselation, point selection and coloring techniques were used for the research, and which technique was better at which point and the advantages it provided were investigated. Triangulation is one of the main topics in computational geometry and it is commonly used in a large set of applications, such as computer graphics, scientific visualization, robotics and image synthesis, as well as in mathematical and natural science. The importance of image storage is increasing day by day. Semiconductor part producers, insurance companies, businesses and law industries, online shopping services require to store massive amounts of product photos. A great deal of medical images have to be stored by healthcare industry. Emerging technologies like autonomous cars and genomics, use immense numbers of image files. Looking at previous studies, there is very little research on the size advantages of using channels in triangulation. Image channels are generally used in areas such as improving underwater photographs, identifying disease in computer aided diagnosis, and cryptography, but the advantages of transmitting and storing image data have not been adequately investigated. In our research, it has been shown experimentally that instead of keeping the vertex coordinates and color of all triangles forming the triangulation, it is sufficient to keep one-third of it, and it is more advantageous in terms of sizing to keep the color of a certain number of clusters instead of keeping the colors of all triangles.