Improving Self-Quotient Image Method of NPR
Authors:
- Huiqin WangHWHuiqin Wang
- Jingchao ChenJCJingchao Chen
- 1 December 2008
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 6, 213-216
Abstract
Self-quotient image (SQI) method was recently proposed, which is applicable to non-photorealistic rendering (NPR). In this paper, we analyse the deficiencies of the original SQI method in generating sketching style from various images and make some improvements on it. The basic idea is to enhance features and weaken noise by enlarging the differences of feature pixels¿ gray values between anisotropic filtrated and isotropic filtrated images and reducing the differences in other areas. The experiment results show that using the improved SQI method, we can obtain sketching style with stronger features and less noises from original photos regardless of their smoothness or noise.Keywords
This publication has 9 references indexed in Scilit:
- Sketching shiny surfacesACM Transactions on Applied Perception, 2006
- More optimal strokes for NPR sketching2005
- SMARTPAPER: An Interactive and User Friendly Sketching SystemComputer Graphics Forum, 2004
- Face recognition under varying lighting conditions using self quotient image2004
- The quotient image: class-based re-rendering and recognition with varying illuminationsIEEE Transactions on Pattern Analysis and Machine Intelligence, 2001
- Geometric Diffusion of Scalar Images2001
- Painterly rendering for video and interaction2000
- Behavioral analysis of anisotropic diffusion in image processingIEEE Transactions on Image Processing, 1996
- Nonlinear total variation based noise removal algorithmsPhysica D: Nonlinear Phenomena, 1992