پیوندهای مفید
آمار
| تعداد نشریات | 20 |
| تعداد شماره‌ها | 1,184 |
| تعداد مقالات | 10,800 |
| تعداد مشاهده مقاله | 72,695,067 |
| تعداد دریافت فایل اصل مقاله | 12,814,677 |
DBCD-MIT: A Dynamic model for Breast Cancer Detection using Multi Inputs Thermograms | ||
| Health Management & Information Science | ||
| مقالات آماده انتشار، پذیرفته شده، انتشار آنلاین از تاریخ 12 دی 1403 | ||
| نوع مقاله: Original Article | ||
| شناسه دیجیتال (DOI): 10.30476/jhmi.2025.105355.1260 | ||
| نویسندگان | ||
| Mahsa Ensafi1؛ Mohammad Reza Keyvanpour* 2؛ Seyed Vahab Shojaedini3 | ||
| 1Faculty of Engineering and Technology, Alzahra University, Tehran, Iran | ||
| 2Faculty of Engineering and Technology, Alzahra University, Tehran, Iran | ||
| 3Department of Biomedical Engineering, Iranian Research Organization for Science and Technology, Tehran, Iran | ||
| چکیده | ||
| Introduction: Breast cancer mortality rates could be greatly reduced with early detection. In the past decade, thermal imaging has been introduced as a way to detect this type of cancer. Although non-invasive, painless, and low-cost, the images resulting from this method require machine learning methods for analysis. Methods: It is common for artificial intelligence systems to analyze thermograms based on a single input of thermography images, or to feed thermograms from different views to a single input detection model, without any processing to extract dependencies and remove redundancies. Using multi-input thermography images in different views correlated with effective correlated information, this article presents a method for detection of breast cancer. To estimate dependencies between thermograms in parallel with remove redundancy, filters are used that are dynamically and separately made for different views. Results: Finally, fusing the information based on their effectiveness in improving the performance of the deep neural network may increase the accuracy of the proposed method in diagnosing the occurrence of breast cancer. In comparison with existing methods, the proposed mechanism can significantly increase breast cancer diagnosis. Experimental results showed that the proposed procedure improved sensitivity and specificity by 1-14 percent and 1-25 percent, respectively, compared to either deep learning or handcrafted approaches. Conclusion: In view of the fact that thermographic images are usually taken from multiple views and the fact that dynamic filters enhance the information extracted from different views, the module presented in this article is an appropriate component for thermogram-based breast cancer classifiers. | ||
| کلیدواژه‌ها | ||
| Thermography؛ Breast cancer؛ Deep learning؛ Diverse perspectives؛ Adaptive filtering | ||
|
آمار تعداد مشاهده مقاله: 5 |
||