# Skin Disease Detector

  

  

This npm package contacts a backend server, (hosted on Heroku: https://ai-skin-server.herokuapp.com/url is where it makes POST requests).
Supports png and jpg file formats.
Chooses between 16 skin conditions (listed below), or `no-conditions`

Full docs:
https://joshyzou.gitbook.io/skinapi/ 
  
## Usage

```
const skinAi = require("@joshyzou/skinconditiondetector")

skinAi("https://img.webmd.com/dtmcms/live/webmd/consumer_assets/site_images/articles/health_tools/lupus_overview_slideshow/dermnet_rf_photo_of_butterfly_rash.jpg")
    .then(res => res.json())
    .then(json => console.log(json));

  

//will return an object like this: 
{ status: 'ok', type: 'lupus' confidence: 1}


```
## Conditions and Accuracy
|Condition|Accuracy  |-|Condition|Accuracy|
|--|--|--|--|--|
| Acne | 95% ||Psoriasis|75%|
|Actinic Keratosis |95%||Ringworm|83%|
|Basal Skin Carcinoma|93%||Rosacea|89%|
|Blisters|70%||Vitiligo|87%|
|Cellulitis|95%|
|Chicken Pox|93%|
|Cold Sore|74%|
|Keratosis Pilaris|91%
|Lupus|81%|
|Measles|81%|
|Melanoma|95%|
|Melasma|79%|