Machine Learning in Medicine:(Skin Cancer Treatment)

What Is Machine Learning?

Machine Learning is a collection of methods whose purpose is to provide software with the ability to learn. It is a type of “deep learning” that allows machines to process information for themselves. On a very sophisticated level, allowing them to perform complex functions like facial recognition. It is a subdomain of artificial intelligence.

How Can Machine Learning Help In Skin Cancer Treatment?

The software named convolutional neural networks(CNN) are better software to detect skin cancer as compared to an expert dermatologist. This is been proven in a newspaper which has published in the leading cancer research journal Annals of Oncology in May 2018. In this journal, authors show the results of a study conducted by researchers in Germany, the US, and France who trained a CNN to identify skin cancer.

At first, the researchers trained a neural network using 100,000 images of malignant melanomas (the most lethal form of skin cancer), together with pictures of benign moles. After the network was trained, they began to compare its performance with the work of 58 dermatologists from 17 countries around the world. And finally, the network detected more melanomas that the expert professor. Not only this but it also misdiagnosed benign moles as malignant less that expert dermatologist.

Technologies for spotting skin cancer

  1. Image Processing Systems

The image processing system works on several steps. At first, it pre-processes the image. In this step, it resizes images and adjusts other aspects such as contrast or brightness. After that, it works in Image Segmentation. In this step, it segments images using tools such as binary masks and edge detection. After that, it extracts the Feature. In this step, it extracts geometric properties of segmented lesions(Area, Perimeter, Greatest Diameter, Circularity Index, Irregularity Index). And finally, the last step is Classification. In this step, it uses k-nearest neighbor algorithms (k-NN) and supports vector machines (SVM) to enable image classification. Image pre-processing. After that, all image are combined with classical machine learning algorithms(k-NN and SVM). Both can be used as binary classification tools (malignant or benign in terms of skin cancer).

2. Hybrid Methods

Researchers have also developed hybrid methods. In the pre-processing workflow, they combine various other machine learning techniques like feature extraction or dimensionality reduction. Beside image processing and features engineering, they also used a feed-forward artificial neural network (ANN) combined with classic machine learning classifiers(k-NN). ANN and k-NN classifier has use in tandem to compute the result. Thats why this is khnown as a hybrid method. This approach allowed them to reach a smashing 95-98% accuracy in detecting cancerous changes.

3. Alternative Approaches

In the latest machine learning technology, researchers have used RBF (Radial Basis Function) and SOM (Self Organizing Maps) to reach 96.15% and 95.45% accuracy in detecting skin cancer changes. The author proposed an original approach to recognize and predict different types of skin cancer. The classification systemhas supervise following the predefined classes of skin cancer. The combination of SOM and RBF for recognition and diagnosis of skin cancer has shown to be more efficient as compared to KNN, Naïve Bayes, and ANN classifiers. The tool achieved the best classification accuracy: 88%, 96.15% and 95.45% for basal cell carcinoma, melanoma, and squamous cell carcinoma, respectively.

Leave a Reply

Your email address will not be published. Required fields are marked *