AI-Based Screening Method for Early Identification of Invasive Ductal Carcinoma in Breast Cancer
The objective of this work is to develop an efficient technique for the early recognition of breast cancer through precise and rapid screening, leveraging artificial intelligence (AI) in the medical field to minimize human error and enhance the likelihood of early intervention, thus potentially incr...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Bilijipub publisher
2024-06-01
|
| Series: | Advances in Engineering and Intelligence Systems |
| Subjects: | |
| Online Access: | https://aeis.bilijipub.com/article_199133_2faf6ab1c984d1dbf462e86b362e7607.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The objective of this work is to develop an efficient technique for the early recognition of breast cancer through precise and rapid screening, leveraging artificial intelligence (AI) in the medical field to minimize human error and enhance the likelihood of early intervention, thus potentially increasing life expectancy and reducing mortality rates. To achieve this, we utilized a deep learning neural network algorithm, employing histopathological microscopic datasets and histological microscopic images from 124 and 576 patients with ductal carcinoma of the breast, respectively. The methodology involved several steps. First, we conducted data preprocessing and image enhancement to improve image quality. Subsequently, we employed the U-Net network for image segmentation to distinguish cancer cells from healthy breast tissue and eliminate outlier data. Next, by leveraging deep neural networks, we extracted effective features, and through a majority vote method, we performed data classification to establish a screening structure for the diagnosis of invasive ductal carcinoma of breast tumors. Our proposed system demonstrated superior performance by achieving 92.8% and 94.4% accuracy, 96% and 93% sensitivity, 91.5% and 92.0% precision, and 98.7% and 96.7% Area Under the Curve (AUC) in two distinct datasets, with minimal errors and high detection speed. This research distinguishes itself by its ability to extract high-level features and provide robust performance in breast cancer diagnosis and classification compared to existing studies. |
|---|---|
| ISSN: | 2821-0263 |