2-D Hybrid Transform and Feature Extracting Algorithm of Fatty Liver Ultrasonic Images

In the fatty, liver ultrasonic images, the desired signal mixed with system noise in spatial domain and frequency domain, and the system noise belongs to a multiplication interference, which make it difficult to extract the fatty liver scatters from a given ultrasonic image. However, the dimensions and the density of the scatters are vital for determining the degree of the fatty liver. To solve the problem, we develop an algorithm based on hybrid 2-D transform, discrete Fourier transform-discrete wavelet transform (2-D DFT-DWT), to extract the liver scatters in the liver scatters from different directions in logarithm-frequency domain. Real processing experiments fir the fatty liver ultrasound images verify, the proposed feature extracting algorithm to be valid.