Adaptive Sound Speed Compensation for Reflection Imaging of Ultrasound Computed Tomography

The reflection image of Ultrasound Computed Tomography (USCT) can provide the breast anatomical structure information and assist the screening and early diagnosis of breast cancer. Because of the individual differences in breast tissues, the USCT reflection image reconstruction may need different sound speeds. To address this problem, we propose an adaptive sound speed compensation strategy. The average breast sound speed of each slice is extracted through one prescan process and then used for the reflection imaging of formal scan. In the prescan process, the signal transmits through the coupling agent only and the signal transmits through the coupling agent and the breast are selected first. Then the maximum value-based time of flights (TOFs) extraction method is used to obtain the corresponding sound speeds, respectively. After that, the average sound speeds of the coupling agent and the breasts can be calculated as the thickness of the coupling agent is known. Finally, the average breast sound speed of each slice is utilized for the reflection imaging of formal scan. The proposed method was tested on a breast phantom, and the influence of sound speed errors on reflection imaging were analyzed. Image clarity evaluation functions, such as Tenengrad gradient function, Brenner gradient function, and Discrete Cosine Transform (DCT) function were used to analyze the imaging results of one dense breast and one fatty breast under the empirical sound speed (1540m/s) and the calculated sound speed. The calculated sound speeds of the fatty and dense breast were 1470m/s and 1516m/s, respectively. The normalized Tenengrad, Brenner gradient values and DCT values of the fatty and dense breast, under the calculated sound speed were about 56%, 63%, 25%, and 16%, 15%, 2% higher than the ones under the empirical sound speed.