Cancer is the leading cause of death in the world and delayed detection being the cause of the most significant factor for its high mortality rate. Computers can help radiologists in analyzing medical images and detection of cancer. Radiomics refers to the computerized extraction of information from medical images and provides the potential for making cancer screening with high rapid and more accurate using machine learning algorithms. Endoscopic imaging and X-ray imaging (Computed tomography) are two common methods used in medical imaging. In this paper, the advantages and limitations of endoscopic and CT scan images discussed. Then the features that can be extracted from endoscopic and CT scans are discussed and finally these two imaging methods are considered and compared to use for computer-aided detection systems.
The paper describes the implementation and testing of two synthetic aperture radar processing algorithms on Jetson TX1 platform using the ArrayFire parallel library. Their execution time was compared when using Jetson’s multicore CPU and GPU. Comparison was also done with a desktop computer. It was discovered that GPU computing on Jetson is faster for both algorithms compared to its CPU. However, the results for a desktop computer depend on the type of an algorithm.
1 June 2019 in Journal of Physics: Conference Series
The automatic detection methods of mobile objects on moving star background in the presence of high-speed image blurring are considered. The approach based on convolutional neural network YOLO v2 for detection and classification of two kinds of objects images is analysed. Estimations of detection accuracy and speed for the approach are based on convolutional neural network of a simplified structure, indicating further prospects for use of such networks.