研究主題一
Topic one
Breast cancer is the most common cancer among women in Taiwan, and the number of breast cancer cases reported annually continues to increase. In 2018, breast cancer ranked fourth in terms of mortality. Early stages (stages 0–2) of malignant breast lesions can be diagnosed during regular screening, and early treatment via advanced medical therapies increases survival rates. Ultrasound imaging, including acoustic radiation force impulse (ARFI) imaging, is the first-line examination technique used to locate breast lesion tissue, which can then be quantitated by virtual touch tissue imaging (VTI). ARFI-VTI elastography is a breast imaging modality that creates two-dimensional (2D) images to visualize the texture details, elasticity, and morphological features of a region of interest (ROI). The 2D Harris corner convolution is applied during digital imaging to remove speckle noise and enhance the ARFI-VTI images for extrapolation of lesion tissue in a ROI. Then, 2D Harris corner convolution, maximum pooling, and random decision forests (RDF) are integrated into a machine vision classifier to screen subjects with benign or malignant tumors. A total of 320 ARFI-VTI images were collected for experiments. In training stages, 122 images were randomly selected to train the RDF-based classifiers and the remaining images were randomly selected for performance evaluation via cross-validation in recalling stages. In a 10-fold cross-validation, promising results with mean sensitivity, mean specificity, and mean accuracy of 86.02%, 87.63%, and 86.97%, respectively, are achieved for quantifying the performance of the proposed classifier. Breast tumors visualized on ARFI-VTI images can be used for rapid screening of malignant or benign lesions by using the proposed machine vision classifier.
研究主題二
Topic Two
High-frequency ultrasound with contrast agents provides contrast enhancement for imaging, has potential for application to drug delivery, and enables local genomics research. However, high-frequency ultrasound has a significant limitation, the achievable signal-to-noise ratio (SNR) and depth of penetration. In this study, we developed a new ultrasound system that involves chirp-coded-excitation ultrasound imaging with chirped pulses as trigger signals and a cardinal frequency of 30 MHz. A chirp is a coded signal that linearly spans a frequency bandwidth B = f2−f1, where f1 and f2 are the starting and ending frequencies, respectively. A new chirped pulses with contrast agents to minimize the attenuation of energy in human tissues increased the signal-to-noise ratio (SNR) by 20 dB for high-frequency ultrasonic flow imaging of the heart of zebrafish and increased the penetration depth to 2.2 mm with pulse compression and a handmade expander. On the other hand, in a microbubble experiment, adopting various echo signal concentrations resulted in the desired distribution of different types of microbubbles. Using the system we developed, we experimentally demonstrated that the chirp-coded excitation reduces the SNR by about 43 dB compared with unipolar and bipolar pulse excitations.
研究主題三
Topic Three
Digitalizing medical images, such as images in ultrasonography or mammography, or magnetic resonance imaging, can be applied for telemedicine applications in telediagnosis and telesurgery, and can be stored in a cloud database via computer networking transmission or wireless communications. Besides, these images contain the patient privacy information. Thus, their reliability and availability should be protected to ensure medical image infosecurity in public channels or open spaces. Medical images can also be hacked by unauthorized people. Therefore, in the picture archiving and communication system (PACS), this study proposes against-hacker attacks with two-round symmetric cryptography models for medical image infosecurity. Hash transformation with multi secret keys is performed to change the pixel values and produce dynamic errors for the two-round encryption processes. In image decryption, two-round decryption processes are employed to estimate the possible hacker attacks at the routing path and to determine the decryption key parameter by using an optimization-based controller. For a case study of mammographic images consisting of 50 benign tumors and 50 malignant tumors, the peak signal-to-noise ratio (PSNR) is employed to evaluate the decryption quality between the plain and decrypted images.
研究主題四
Topic Four
Peripheral arterial disease (PAD) is highly prevalent in haemodialysis (HD) patients with type 2 diabetes. Atherosclerosis may occur in both lower and upper peripheral arteries, causing progressive dialysis access stenosis in HD patients. To assess the risk of PAD, non-invasive bilateral photoplethysmography (PPG) can be used to obtain continuous variations in blood flow volume in in vivo examinations. The authors propose an astable multivibrator to model the peripheral circulation system and to produce PPG oscillation with time constants, duty ratio (rising time), and amplitude ratio of systolic and diastolic pressures. Then, the bilateral differences in the time constant and duty ratio are used to separate the normal condition from PAD. The machine learning decision-making process utilises a screening method to automatically detect subjects with and without the risk of PAD. The radial-based function is employed to parameterise the similarity and dissimilarity levels using probability values. Colour relation analysis incorporates the probability values into the perceptual colour relationships for PAD screening. The experimental results indicate that in comparison with bilateral timing parameters, degree of stenosis, and resistive index, the proposed screening method is efficient in preventing complications of PAD and is easily implemented in an embedded system.