Skin diseases detector

• 

  The output is that the patience has Eczema and is advised to schedule an appointment with a doctor

• 

  This is an image that I took for an example that this baby has Eczema and the model detected Eczema

Inspiration

We mentioned usually have uncommon skin structures but they do not know what it is, for example, when people got sunburned and they have something appears on their skin and they do not know what unprescribed treatment should they purchased. Therefore, I want to have an application that can identify the diseases through the pictures of the symptoms.

What it does

This deep learning model helps to identify uncommon structures on skin and then it can come up with unprescribed medications for unharmful diseases and for harmful diseases like when it detect a symptom of skin cancer, it will advise to schedule a meeting with a doctor as soon as possible.

How we built it

We used ResNet architecture and tensorflow to train.

Challenges we ran into

Our training took more than 12 hours and one more challenge is that we cannot connect tensorflow to use the GPU on our laptop to accelerate the speed of training process.

Accomplishments that we're proud of

We are proud of our model's accuracy is over 70% and we can learn a little bit more about the medical field.

What we learned

We have more knowledge in the medical field, which helps us to have more life experiences, and deep learning.

What's next for Skin diseases detector

We will make the model more accurate using Faster RCNN, which have more technologies to identify faster and more accurate

Built With

• python
• resnet
• tensorflow