Study of the performance change in digital mammography systems depending on the total number of examinations.

Digital mammography equipment should operate at a high-performance level for detecting breast cancer over the lifetime of the systems. This study investigates the performance of the Fujifilm Amulet Innovality full-field digital mammography system in relation to the number of exposures. The performance of twelve systems, from new and up to 17 000 examinations, was compared. The x-ray output, half-value layer (HVL), contrast-detail (C-D) detectability, contrast-to-noise ratio (CNR), mean glandular dose (MGD), signal transfer property (STP), pre-sampled modulation transfer function (MTF), normalized noise power spectra (NNPS), detective quantum efficiency (DQE) tests were performed to determine the performance change. The noise sources were also analyzed in the spatial domain. The x-ray output and HVL values at 28 kV with tungsten anode and rhodium filter were 15µGy/mAs ± 1.0 (coefficient of variation (cov) = 7.0%) and 0.54 mmAl ± 0.008 (cov = 1.5%), respectively. The average MGDs for 60 mm equivalent breast to reach the achievable image quality level for 0.1 mm and 0.25 mm detail diameters of circular discs on the CDMAM 3.4 phantom image were 1.17 mGy ± 0.13 (cov = 11%) and 1.40 mGy ± 0.09 (cov = 6%), respectively. The average MGD for 53 mm equivalent breast was 1.08 mGy ± 0.14 (cov = 13%) at dose setting Normal in automatic exposure control (AEC) mode. All tested devices demonstrated good linearity with R2 ≥ 0.999 in the STP curves. The average MTF at 5.0 mm-1spatial frequency was 0.66 ± 0.007 (cov = 1.0%). The average NNPS was 2.38 × 10-6mm2 ± 2.13 × 10-7(cov = 9.0%) at 5.0 mm-1. The average DQE value at 5.0 mm-1was 0.33 ± 0.02 (cov = 5.0%), and the mean peak DQE was 0.66 ± 0.03 (cov = 5.0%). The mean power coefficient (b), determined from the power relationship between linearised standard deviation and detector dose while analyzing noise sources, was 0.49 ± 0.007 (cov = 1.4%). There was no gradual change in the x-ray tube and detector performance by the number of exposures, and taking up to 68 000 images did not decrease the quality of the systems.

Detection of COVID-19 and its pulmonary stage using Bayesian hyperparameter optimization and deep feature selection methods.

Since the first case of COVID-19 was reported in December 2019, many studies have been carried out on artificial intelligence for the rapid diagnosis of the disease to support health services. Therefore, in this study, we present a powerful approach to detect COVID-19 and COVID-19 findings from computed tomography images using pre-trained models using two different datasets. COVID-19, influenza A (H1N1) pneumonia, bacterial pneumonia and healthy lung image classes were used in the first dataset. Consolidation, crazy-paving pattern, ground-glass opacity, ground-glass opacity and consolidation, ground-glass opacity and nodule classes were used in the second dataset. The study consists of four steps. In the first two steps, distinctive features were extracted from the final layers of the pre-trained ShuffleNet, GoogLeNet and MobileNetV2 models trained with the datasets. In the next steps, the most relevant features were selected from the models using the Sine-Cosine optimization algorithm. Then, the hyperparameters of the Support Vector Machines were optimized with the Bayesian optimization algorithm and used to reclassify the feature subset that achieved the highest accuracy in the third step. The overall accuracy obtained for the first and second datasets is 99.46% and 99.82%, respectively. Finally, the performance of the results visualized with Occlusion Sensitivity Maps was compared with Gradient-weighted class activation mapping. The approach proposed in this paper outperformed other methods in detecting COVID-19 from multiclass viral pneumonia. Moreover, detecting the stages of COVID-19 in the lungs was an innovative and successful approach.