Advancements in AI for Early Breast Cancer Detection
One of the most common types of cancer found in females is breast cancer, with 287,500 cases diagnosed in 2022, according to a recent study, “Classification of Breast Cancer Using Transfer Learning and Advanced AI-Biruni Earth Radius Optimization” by Alhussan et al. Out of that population subset, around 43,250 people have died. Despite that, breast cancer is significantly easier to treat when diagnosed in earlier stages; however, a simple mammogram is not always effective at detecting early-stage cancers. A mammogram uses a machine learning algorithm trained to look for specific patterns and anything in the body that might be associated with the cancer, such as lumps, masses, or cysts, all classified as nodules. Essentially, it's a special X-ray machine that images your breast when placed upon it. A drawback of this screening method is that it uses small doses of radiation which can be harmful to one's health. In most cases, breast cancer is diagnosed in further stages which often makes it more difficult to treat because the cancer is more progressed and may become harder to resect.
According to a new article, researchers are attempting to find a way to detect breast cancer in earlier stages using a deep-learning artificial intelligence method. In essence, deep learning modulates a human brain's complex thinking patterns via artificial intelligence to gain better images and more precise results.
The two artificial intelligence programs with the best accuracy were ResNet50 and MobileNet, with 78.4% and 74.3% accuracy, respectively. Using certain malignancy indicators — signs that can help to find a tumor — such as the area and radius of the tumor allowed the researchers to help the artificial intelligence detect the cancerous areas. They came up with a method to identify the breast cancer abnormalities by using a properly improved image which allowed for a 92% precision rate.
According to a new article, researchers are attempting to find a way to detect breast cancer in earlier stages using a deep-learning artificial intelligence method. In essence, deep learning modulates a human brain's complex thinking patterns via artificial intelligence to gain better images and more precise results.
The two artificial intelligence programs with the best accuracy were ResNet50 and MobileNet, with 78.4% and 74.3% accuracy, respectively. Using certain malignancy indicators — signs that can help to find a tumor — such as the area and radius of the tumor allowed the researchers to help the artificial intelligence detect the cancerous areas. They came up with a method to identify the breast cancer abnormalities by using a properly improved image which allowed for a 92% precision rate.
The world of artificial intelligence is colliding with medicine to help make advances in the field.
Image Source: redgreystock
Another artificial intelligence used in this process is the AlexNet deep network which uses a technique known as transferring learning. Transfer learning is a type of machine learning that allows the reuse of a pre-trained model when encountering new problems. The AlexNet was originally used for image recognition and classification tasks, but it has been further developed to help in the early detection of breast cancer using the results from mammogram imaging. After using several different networks, they worked towards making an optimized neural network model called the Advanced AI-Biruni Earth Radius optimization algorithm, or ABER for short. With this, the researchers were able to produce a classification accuracy of 97.95%.
While the study has made some advancements in early cancer detection, it is important to note that it is still being researched and may take several years until it is fully available and incorporated into modern treatments. However, this increases the prospects for new artificial intelligence models to be integrated into the field of medicine and with a high potential to change patient care.
While the study has made some advancements in early cancer detection, it is important to note that it is still being researched and may take several years until it is fully available and incorporated into modern treatments. However, this increases the prospects for new artificial intelligence models to be integrated into the field of medicine and with a high potential to change patient care.
Featured Image Source: Klaus Nielsen
RELATED ARTICLES
|
Vertical Divider
|
Vertical Divider
|
Vertical Divider
|
