Detection Of Mammographic Lesions Research Articles

Computer-aided diagnosis for breast cancer detection and classification using optimal region growing segmentation with MobileNet model

Globally, breast cancer is considered a major reason for women’s morality. Earlier and accurate identification of breast cancer is essential to increase survival rates. Therefore, computer-aided diagnosis (CAD) models are developed to help radiologists in the detection of mammographic lesions. Presently, machine-learning (ML) and deep-learning (DL) models are widely employed in the disease diagnostic process. In this view, this paper designs a novel CAD using optimal region growing segmentation with a MobileNet (CAD-ORGSMN) model for breast cancer identification and classification. The proposed CAD-ORGSMN model involves different stages of operations, namely, pre-processing, segmentation, feature extraction, and classification. Primarily, the proposed model uses a Weiner filtering (WF)–based pre-processing technique to remove the existence of noise in the mammogram images. The CAD-ORGSMN model involves a glowworm swarm optimization (GSO)–based region growing technique for image segmentation where the initial seed points and threshold values are optimally created by the GSO algorithm. Besides, a MobileNet-based feature extractor is used in which the hyperparameters of the MobileNet model are optimally selected using a swallow swarm optimization (SSO) algorithm. Lastly, variational autoencoder is applied as a classifier to determine the class labels for the input mammogram images. The utilization of the GSO algorithm for the region growing technique and the SSO algorithm for hyperparameter optimization helps to considerably improve the breast cancer detection performance of the CAD-ORGSMN model. The performance validation of the CAD-ORGSMN model takes place against the Mini-MIAS database, and the obtained results highlighted the promising performance of the CAD-ORGSMN model over the recent state-of-the-art methods in terms of different measures.

Computer Aided Detection of Mammographic Lesions Using Deep Learning

Computer Aided Detection of Mammographic Lesions Using Deep Learning

Novel Segmentation Technique Using Wavelet Based Active Contour Model for Detection of Mammographic Lesions

Novel Segmentation Technique Using Wavelet Based Active Contour Model for Detection of Mammographic Lesions

Effect of dose reduction on the detection of mammographic lesions: A mathematical observer model analysis

The effect of reduction in dose levels normally used in mammographic screening procedures on the detection of breast lesions were analyzed. Four types of breast lesions were simulated and inserted into clinically-acquired digital mammograms. Dose reduction by 50% and 75% of the original clinically-relevant exposure levels were simulated by adding corresponding simulated noise into the original mammograms. The mammograms were converted into luminance values corresponding to those displayed on a clinical soft-copy display station and subsequently analyzed by Laguerre-Gauss and Gabor channelized Hotelling observer models for differences in detectability performance with reduction in radiation dose. Performance was measured under a signal known exactly but variable detection task paradigm in terms of receiver operating characteristics (ROC) curves and area under the ROC curves. The results suggested that luminance mapping of digital mammograms affects performance of model observers. Reduction in dose levels by 50% lowered the detectability of masses with borderline statistical significance. Dose reduction did not have a statistically significant effect on detection of microcalcifications. The model results indicate that there is room for optimization of dose level in mammographic screening procedures.

Digital Mammography: Comparative Performance of Color LCD and Monochrome CRT Displays

To evaluate the comparative performance of high-fidelity liquid crystal display (LCD) and cathode ray tube (CRT) devices for mammography applications, and to assess the impact of LCD viewing angle on detection accuracy. Ninety 1 k x 1 k images were selected from a database of digital mammograms: 30 without any abnormality present, 30 with subtle masses, and 30 with subtle microcalcifications. The images were used with waived informed consent, Health Insurance Portability and Accountability Act compliance, and Institutional Review Board approval. With postprocessing presentation identical to those of the commercial mammography system used, 1 k x 1 k sections of images were viewed on a monochrome CRT and a color LCD in native grayscale, and with a grayscale representative of images viewed from a 30 degrees or 50 degrees off-normal viewing angle. Randomized images were independently scored by four experienced breast radiologists for the presence of lesions using a 0-100 grading scale. To compare diagnostic performance of the display modes, observer scores were analyzed using receiver operating characteristic (ROC) and analysis of variance. For masses and microcalcifications, the detection rate in terms of the area under the ROC curve (A(z)) showed a 2% increase and a 4% decrease from CRT to LCD, respectively. However, differences were not statistically significant (P > .05). The viewing angle data showed better microcalcification detection but lower mass detection at 30 degrees viewing orientation. The overall results varied notably from observer to observer yielding no statistically discernible trends across all observers, suggesting that within the 0-50 degrees viewing angle range and in a controlled observer experiment, the variation in the contrast response of the LCD has little or no impact on the detection of mammographic lesions. Although CRTs and LCDs differ in terms of angular response, resolution, noise, and color, these characteristics seem to have little influence on the detection of mammographic lesions. The results suggest comparable performance in clinical applications of the two devices.