Facts and research state that
- Skin cancer is the most common cancer in the United States.
- Current estimates are that one in five Americans will develop skin cancer in their lifetime.
- It is estimated that approximately 9,500 people in the U.S. are diagnosed with skin cancer every day
- Nearly 20 Americans die from melanoma every day
- The annual cost of treating nonmelanoma skin cancer in the U.S. is estimated at $4.8 billion, while the average annual cost of treating melanoma is estimated at $3.3 billion
- Basal cell and squamous cell carcinomas, the two most common forms of skin cancer, are highly curable if detected early and treated properly
- The five-year survival rate for people whose melanoma is detected and treated before it spreads to the lymph nodes is 99 percent.
If we can detect the type of skin cancer in early stages, we can save almost every victim. This cancer classification model can help healthcare organizations especially in developing countries to improve their services and reduce their costs for the patients.
What it does
We created a Machine learning model(Convolutional neural network) using TensorFlow 2.0 and python to detect the seven major types of skin cancer namely
- Actinic keratoses and intraepithelial carcinoma / Bowen's disease (akiec)
- Basal cell carcinoma (bcc)
- Benign keratosis-like lesions (solar lentigines / seborrheic keratoses and lichen-planus like keratosis), (bkl)
- Dermatofibroma (df)
- Melanoma (mel)
- Melanocytic nevi (nv)
- Vascular lesions (angiomas, angiokeratomas, pyogenic granulomas and hemorrhage, vasc).
How we built it
- We selected the skin cancer dataset from Harvard Dataverse (https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/DBW86T)
- Based on the metadata, we separated the 10015 image files into sub categories and created training and testing datasets.
- Uploaded the dataset into Google drive so that it can be accessed in the Colab environment
- Created the CNN model using Tensorflow 2.0
Challenges we ran into
Learning TensorFlow and understanding the differences between TensorFlow and TensorFlow 2.0 was a major task for us. we are also new to google colab and learning how to use it took some time.
Accomplishments that we're proud of
Learning TensorFlow and creating a model that can benefit and bring about a positive change in the world for many people around the world on this open-source platform
What we learned
We learnt about types of skin cancer and its preventive measures, also using TensorFlow and google colab
What's next for Skin Cancer Classification
we are planning to use TensorflowJS and integrate our model to a nodeJS based website and launch it online so that everyone can access this. The future scope of this cancer classification project would be the improvement of the model performance and applications that would benefit healthcare professionals and creating similar models to diagnose or classify diseases from X-rays, CT scan and MRI images