Program Information
Full-Reference Image Quality Assessment-Based B-Mode Ultrasound Image Similarity Measurement
Kele Xu1*, Zhifeng Gao2 , Hengxing Cai3 , (1) National University of Defense Technology, Changsha, Hunan, (2) Peking University, Peking, Peking, (3) Sun Yat-Sen University, Guangzhou, Guangdong
Presentations
SU-I-GPD-U-2 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose: Potential applications of the ultrasound image similarity measurement seem evident in several aspects. To name a few, ultrasound imaging quality assessment, abnormal function region detection, etc. A comparative study was made on the full-reference image quality assessment methods for ultrasound image visual structural similarity measure.
Methods: To make a quantitative comparison study between different similarity indices on the ultrasound image similarity measurements, we conducted an experimental study on the synthetic data from the perspectives on the speckle sensitivity and periodical local motion. The similarity indices used for this comparative study include: Mean Squared Error (MSE), Peak Signal-to-Noise Ratio (PSNR), Structural Similarity (SSIM), Multi-Scale Structural Similarity (MS-SSIM), Complex Wavelet Structural Similarity (CW-SSIM) and Visual Information Fidelity (VIF). Moreover, the ultrasound liver sequences are also used to evaluate the performance of different similarity indices.
Results: MSE, PSNR, VIF and SSIM indices cannot provide stable similarity descriptions with the occurrence of the speckle noise. MS-SSIM provides better performance with comparison to SSIM, while CW-SSIM has superior performance. CW-SSIM is able to characterize the periodic motion of the landmarks in the image sequence. The Pearson correlation between the position of the landmarks in the ultrasound image sequences and the similarity index is calculated for different similarity measurements. The CW-SSIM provides highest correlation coefficient (0.97) between the position of the landmark and the similarity index, which demonstrates that CW-SSIM can characterize the periodic motion in the ultrasound image sequences.
Conclusion: The results demonstrate that the CW-SSIM can provide more accurate similarity description between the frames in the ultrasound image sequences.
Contact Email: