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% |