7 December 2012 in Radiological Physics and Technology
1 May 2023 in Scientific Reports
This study proposes a deep convolutional neural network (DCNN) classification for the quality control and validation of breast positioning criteria in mammography. A total of 1631 mediolateral oblique mammographic views were collected from an open database. We designed two main steps for mammographic verification: automated detection of the positioning part and classification of three scales that determine the positioning quality using DCNNs. After acquiring labeled mammograms with three scales visually evaluated based on guidelines, the first step was automatically detecting the region of interest of the subject part by image processing. The next step was classifying mammographic positioning accuracy into three scales using four representative DCNNs. The experimental results showed that the DCNN model achieved the best positioning classification accuracy of 0.7836 using VGG16 in the inframammary fold and a classification accuracy of 0.7278 using Xception in the nipple profile. Furthermore, using the softmax function, the breast positioning criteria could be evaluated quantitatively by presenting the predicted value, which is the probability of determining positioning accuracy. The proposed method can be quantitatively evaluated without the need for an individual qualitative evaluation and has the potential to improve the quality control and validation of breast positioning criteria in mammography.
1 April 2016 in American Journal of Roentgenology
OBJECTIVE. The purpose of this study was to investigate the optimal b values required for the generation of computed high-b-value DW images. SUBJECTS AND METHODS. Brain DWI was performed for eight subjects using 16 b values, ranging from 0 s/mm2 to 5000 s/mm2. The fractional value as the turning point between the fast and slow components was determined using a diffusion decay curve. Image contrast in the white matter, gray matter, thalamus, putamen, and lateral ventricle was compared between the acquired high-b-value images and the computed high-b-value images derived from various b value images. In addition, the signal-to-noise ratio of each computed image and acquired image was measured and compared. RESULTS. The fractional values obtained from the diffusion decay curve were between 1200 and 1800 s/mm2. Image contrast in all regions was nearly equivalent between the acquired and computed high-b-value images when b values higher than the fractional value were used, whereas image contrast was significantly different between the two sets of images when b values lower than the fractional value were used (p < 0.05). The computed images derived from low-b-value images showed a higher signal-to-noise ratio, compared with the acquired images. CONCLUSION. Optimal b values should be considered when acquiring images for the derivation of computed DW images.
30 July 2012 in Hepatology Research
1 January 1991 in Japanese Journal of Medicine
We performed 93 sclerotherapy sessions on liver cirrhosis patients with recurrent variceal bleedings. In each session, hypertonic glucose, thrombin and 1% polidocanol were consecutively injected into the varices, and changes in the hemostatic system were examined in relation to the symptoms observed during the treatment. Patients underwent sclerotherapy with no complaints in 62 (67%) sessions, and complained of slight symptoms of general fatigue and headache in 19 (20%). In the other 12 (13%) sessions, the procedure was discontinued due to marked manifestation of these symptoms. All symptoms were temporary and disappeared completely after the procedure. These temporary symptoms were closely related to changes in coagulation tests similar to those of disseminated intravascular coagulation, which were observed just after the treatment. Possible activation of the renal kallikrein-kinin system following injection sclerotherapy was also demonstrated.
20 March 2020 in Technology and Health Care
BACKGROUND: Applied research on artificial intelligence, mainly in deep learning, is widely performed. If medical images can be evaluated using artificial intelligence, this could substantially improve examination efficiency. OBJECTIVE: We investigat