A very recent report by World Health Organization(WHO) shows that skin diseases are among the most common of all human health afflictions and affect almost 900 million people in the world at any time. Five common conditions account for over 80% of all skin diseases.
There are a lot of doctors and researchers doing their best in this field of medical science, there are Machine Learning models present with very high accuracy and precision to help the researchers in this field to make their study more effective, but all these technological advancements are still quite far away from the reach of the general public.
There are still very less or no technological resources which can add some value to the lives of people who know very little about such disorders.
To tackle this issue, we have created a web-based skin disease predictor which can be used as a native application by the users to predict their skin disorders just by uploading a picture of their skin using a machine learning model trained on the top of TensorflowJS. This app also supports real-time detection of skin diseases by which users can directly use their cameras to focus on their textured areas and get instant results.
Further, this application also recommends nearby doctors automatically by using their current location. This app has the ability to predict over 10 skin diseases quite accurately.
We've also included a feature of global chat within this application to help the users in posting their queries of each kind and get their doubts resolved. This way, users can even take help from real doctors through this feature without even going out.
You can use this link to navigate through the website- Infinity SknCure Website
- Head on to the same website mentioned above on your device.
- In the browser options, search for the 'Add to home screen'/'Install App' option and tap it. (You'll find an install button in the address bar if you're using Chrome on windows/mac).
- Wait for the device to install the app.
- An app with the name Infinity-Skncure will be visible on your home screen.
- Can detect over 11 skin diseases accurately.
- Trained on 300,000 images.
- It's a Web-App and is installable via the browser. It also is platform-independent and can be installed on any device including but not limited to Android, iOS, Windows and macOS.
- Since the Machine Learning model is included in the package, users don't need to be present online to get their results.
- Options to select between 'Import File' and 'Real-Time', Import file lets the users import their previously clicked images and the Real-Time feature is for Real-Time Skin Disease Detection.
- Built on the top of TensorflowJS, ml5.js and p5.js.
- Ability to show users' nearby skin doctors by taking up the users' location and performing a search.
- Interactive User Experience and an easy-to-understand User Interface makes it easier for elderlies to get the results.
- Fast, scaleable and lightweight, works on slow networks too.
- Runs on users' browsers, which means there's no use of any additional or external hardware.
- The model has been trained on over 1500 images which were iterated 200 times each.
- The model is trained keeping in mind the various skin types (normal, dry, oily, combination, etc.), colors (white, black, brown, porcelain, ivory, sand, beige, etc.), body-parts (scalp, forehead, ears, nose, cheeks, lips, chin, neck, chest, back, arms, hands, palms, reproductive organs (both male and female), thighs, feet and toes).
- Acne
- Herpes Zoster
- Corn
- Rosacea
- Psoriasis
- Contact Dermatitis
- Cutaneous Horn
- Eczema
- Melanoma
- Urticaria
Skin diseases are among the most common health problems in humans. Considering their significant impact on the individual, the family, the social life of patients, and their heavy economical burden, the public health importance of these diseases is underappreciated. Cite
According to a survey skin disorders increase with age and are more frequent in men (72.3 percent) than in women (58.0 percent). Nearly two-thirds of the affected people are unaware of their abnormal skin conditions.
With Infinity-SknCure people get an opportunity to self-diagnose their skin disorders from the comfort of their home. It offers an opportunity to screen people and to identify multiple conditions on a single platform. An integrated approach in communities and schools can potentially reduce costs and cut down delays in diagnosis as well as promote skin health for all.
Reports also point that men are very less likely to use sunscreens which is a major reason why many people suffer from skin cancers and it again points towards the associated stigmas and with this application, we want to fight these social stigmas.
- US National Library of Medicine
- Science Daily Journals
- WHO neglected diseases
- National Center for Biotechnology Information
- A great application to serve the underprivileged group with a nice implementation.
~ Dr S.C. Jain (Professor, Computer Science Department, Rajasthan Technical University, Kota)
We're looking forward to expanding the model to cover more diseases for the prediction. We're also working towards making it more user-friendly by adding various functions some of which are listed below -
- Making chat feature more user-friendly by providing the feature of replying in threads.
- Providing in-app tips to users for their respective identified disease.
- Enabling the users to sign in and measure the progress they've made towards healthy skin.
- Dermnet NZ - https://dermnetnz.org
- DermIS - https://dermis.net
- Dermatology Atlas - http://www.atlasdermatologico.com.br/browse.jsf