A New Homomorphic Filtering Based on Wavelet Transform for Endoscope Image Enhancement
Yuhua Zhang · Tance yu kongzhi xuebao · 2008
Because of the variety of light resource intensity,endoscope image suffers from undesired luminance distribution and loss of contrast.The illumination-reflectance model and homomorphic filtering are discussed for endoscope image contrast enhancement.A novel homomorphic filtering based on wavelet transform is proposed,which adopts fast 2D wavelet transform instead of Fourier transform.The wavelet decomposition coefficients are processed by using nonlinear mapping to attenuate the low spatial frequencies and amplify the high spatial frequencies.The experiments show that the new method can eliminate the non-uniformity luminance distribution of endoscope image effectively.Moreover,the method performs well in enhancing the local contrast of endoscope image,and its performance is better than the traditional homomorphic filtering.