Cancer impacted about nearly 10 million deaths in the world (WHO) by 2020. One of the most common types of cancer was skin (non-melanoma) with 1.2 million cases. Skin cancer was common but not everyone was aware of this cancer. Most people not doing self-diagnoses, they might not understand how to do it periodically. Because of the late action, the death rate of these cases is still on top of the chart. Through technology, self diagnosis can be provided in an easier way. This may help to prevent skin cancer and reduce the death rate.
training analysis shows below:
- Training Accuracy
The training shows that accuracy is getting higher and consistent by epoch 20
- Training Loss
The training shows that loss is getting lower and consistent by epoch 20
The Confusion Matrix shows that classification model did a good job. Then, there is one problem where classification
did not show satisfy result and that is NV or Melanocytic Nevi (indicated by number 4).