Abstract:
Breast cancer is the second leading cause of mortality and fatal disease among women, making it a serious public health issue at present. Timely diagnosis plays a pivotal role in improving prognosis and survival rates, underscoring the urgency for early detection. Fortunately, advancements in the analysis of radiographic imaging and histopathological images is feasible to diagnosis and done by experienced radiologist. Recent developments in artificial intelligence (AI), in particular machine learning (ML) and Deep Learning (DL), have demonstrated promising outcomes in various fields, including the internet of things, automated machinery, and healthcare. DL techniques are dominating medical image analysis for the early detection of breast cancers. This study discusses the overall potential of ML and DL techniques to automatically grade, recognize, and assess the abnormal features that will empower radiologists to provide accurate diagnoses and facilitate personalized health care. In this study, a brief overview of the analysis of traditional and advanced techniques was discussed. This study highlights the open issues, research gaps, and future directions for the early detection and diagnosis of breast cancers. In this paper, we proposed an improved DL based classification pipeline for the detection of breast cancer from Whole-slice images. This article presents a comparative analysis of two feature extraction methodologies that have been employed for breast cancer classification: (1) hand-crafted features from the histogram of oriented gradient, local binary pattern, and gray-level co-occurrence matrices; and (2) deep features from deep learning architecture. The retrieved handcrafted features and deep learning features are applied to classification algorithms such as support vector machine, gradient boosting, random forest, XGBoost, and softmax for the identification of breast cancers. It has been found that the proposed architecture obtained excellent results in terms of accuracy, sensitivity, specificity, and F1 score. This study may help clinicians to early progression with high accuracy for timely interventions and utilization of advanced DL technology, with a specific focus on its applications in the detection of breast cancers.
Audience Take Away Notes:
- Analyze and compare various methods for extracting features, such as hand-crafted and deep features, in order to automatically grade and assess features in breast cancer diagnosis.
- Investigating the capabilities of advanced machine learning methods, such as Generative Adversarial Network (GAN) and convolutional neural network (CNN), for detecting breast cancer.
- Various classification techniques, including support vector machine, gradient boosting, random forest, XGBoost, and softmax, are used to detect breast cancer.
- Radiologists may use the insights obtained from this research into their clinical procedures, enhancing the precision and effectiveness of breast cancer screening, thereby helping patients through personalized healthcare solutions.
- Researchers may investigate these approaches to further develop their research in medical imaging and deep learning applications.
