|
|
COMPUTER GRAPHICSRay Tracing
Computer Vision Error DiffusionImage processing methods that are not based on partial differential equations are not as powerful as the PDE-based methods, and therefore, their use is limited. However, due to their speed, they are effective when it is critical to get a result almost instantly (i.e. camera processing of an image, monitor picture conversion.) Here, one of such methods, namely error diffusion, is used to overcome the lack of intrinsic colors by diffusing the error around the neighboring pixels and making the human eye believe that more colors are present than are really available. A 24-bit color image (16 million colors) is represented as a 3-bit color image (8 colors) by using pure thresholding and by error diffusion. Grayscale images can be represented in a similar way. Error diffusion significantly improves the quality of the picture, making it a better visual approximation of the original image.
|
•
Igor Yanovsky |