The Research Of Ultrasound Tissue Characterization Based On Hilbert-huang Transform

Ultrasound tissue characterization(UTC)is one of the most promising technologies for studying tissue microstructure.It complements the shortcomings of traditional medical imaging diagnostic techniques that can only understand the macroscopic anatomy of tissue.Studies in recent decades have shown that in the numerous studies of UTC,spectral analysis based on ultrasonic backscattered signals has shown greater development potential due to its consideration of signal amplitude and spectral information.This dissertation proposes to use ultrasound frequency(RF)time series of the tissue as the signal source and Hilbert-Hua ng transform(HHT)signal time-frequency analysis method to perform study of UTC.The specific work of the dissertation is as follows.1.The HHT-based ultrasound RF time series feature extraction algorithm was proposed.The HHT-based ultrasonic tissue feature extraction system was designed and developed to extract the features of the experimental subjects in the study.The system is also convenient for doctors to extract parameters that can reflect the tissue biological characteristics,and then it guides clinical trials and improves the level of auxiliary diagnosis.2.In order to verify the valid ity of the proposed method and system for UTC,it was used for the early evaluation of tumor chemotherapy.The study was conducted using nude mouse mammary tumors that had not received chemotherapy(control group),1 mg/kg paclitaxe l chemotherapy(Treatment 1),and 3 mg/kg paclitaxel chemotherapy(Treatment 2)as experimental subjects.During the experiment,ultrasound RF signals were collected to construct ultrasound RF time series,and it used the HHT-based feature extraction system to extract features from ultrasound RF time series,and then the intragroup distance in the three groups,the distance between the treatment group 1 and the control group and the distance between treatment 2 groups and the control group were calculated by the extracted features.The experiment results showed that the distance mean and standard deviation between the treatment group 1 and the control group and between the treatment 2 group and the control group were both greater than the distance mean and standard deviation within the control group.At the same time,the distance mean and standard deviation between the treatment 2 group and the control group was greater than the distance mean and standard deviation between the treatment 1 group and the control group.These results indicate that paclitaxel is effective for breast cancer chemotherapy,and is consistent with the results of pathological analysis,and the effect of large doses of chemotherapy is more obvious.Therefore,the HHT-based ultrasound tissue characterization method is effective for the early evaluation of tumor chemotherapy.