A Hybrid Vector Filter for Color Image Noise Reduction

• Main Authors: Shi-Jie Wu, 吳世傑
• Other Authors: Jiunn-Lin Wu 吳俊霖
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/65166481754513662391

碩士 === 國立中興大學 === 資訊科學系所 === 94 === A hybrid vector filter is proposed for noise reduction in color images. By combining the vector median filter based on the impulse noise detection approach and the weight adaptive vector filtering for noise removing and edge preserving, the proposed method achieves the effective noise removal and details preserving. The main advantages of the proposed method include: (1) the impulse noise detection algorithm can reduce the faults of noise detection at the edge and details effectively. Moreover, using the iterative approach, it provides the high accuracy for noise detection; (2) the proposed weight adaptive function obtain more proper weighting value for pixels at the edges. It not only suppresses the Gaussian noise effectively but also preserves the edges and details information in the images. The experimental results show that the proposed method can remove the impulse noise and suppress Gaussian noise effectively.