Skin Cancer Check with Skin Photos
(This software is in experimental stage. Please use with caution.)
This skin cancer check system detects potential skin cancer from skin photo images. It identifies potentially cancerous skin spots, moles, melanoma, etc. This is free to use software powered by AI Machine Learning Deep Neural Network Models. It is highly accurate. But is NOT 100% accurate! (See below Limitations & About Accuracy.) Note that this is not a medical diagnosis system. For accurate medical diagnosis, you may need biopsy and pathological tests. There are limitations what photos can tell you! Certain skin cancer looks alike pimple or rash. This system can make errors. Note that pimple or rash normally improves or disappears in days. If conditions don't improve, please see a doctor or dermatologist for medical diagnosis and treatments.
By using this software, you are agreeing that we do not have any responsibility from software errors. If you agree, then tick the following and proceed;
Does your spot or mole have the following conditions?
Please tick each correctly, then choose a file, then a lesion!
[Instructions] Take color photos of skin and upload into a computer or mobile device. Some mobile devices may allow you to use camera directly. Spots should not be tiny in pictures. Minimum 10% and up to 70% of image size is recommended. For more, please read Photo Requirements. Tick above questions correctly and press the "Choose File" button and select your skin image file. With the mouse, draw a rectangular box around a suspected spot or mole. From the top left corner, drag the mouse to the bottom right corner. Note that selected box should be tight but slightly wider to the suspected spot or mole as shown in the following figures. Otherwise results can be random!
Try multiple times with wider or smaller box areas. When multiple spots are adjoined as one (as in the 3rd above image), try each spot individually. Results will be displayed immediately within a second or two. Results are shown in percentages which indicate skin cancer probabilities, ranging 0% to 100%. Red means highly likely. Yellow is medium likelihood. Green is low probability. Note that malignancy is not tested here. Normally cancer-like moles and spots will have high score. For accurate diagnosis, please check with a doctor or dermatologist! If you have ticked at least one of the conditions, seeing a doctor may be a safer option .
To be accurate, photos require that the following conditions are met. Please take photos with this requirement in mind;
- Color photos. Black and white photos are not supported.
- Relative sizes of skin lesion in photos should not be too tiny. Minimum 10% and up to 70% of photo size is recommended. If relative size is too small, it's difficult to select a lesion accurately and thus can cause errors.
- Photos should not be everything too bright or too dark. Clearly discernable skin spots or moles are recommended.
- Avoid excessive light reflection. It can cause errors.
- Blurry photos are bad. Check camera focus while taking photos.
- Supported file formats: JPEG, PNG, WebP, GIF, BMP, SVG, etc.
[False negatives] Certain skin cancer looks similar to pimples or skin rash. In this case, this system may classify as low risk. Note that pimples and skin rash improve or disappear in days. If conditions don't improve in days, see a doctor or dermatologist.
[False positives] Some bad skin infections look similar to skin cancer. This system may make mistakes. If conditions don't improve, you may need to see a doctor for treatment.
Note that this system doesn't work well with dark skin. This is an inherent limitation of this system.
Three deep neural network models trained with 1 million photo images, derived from about 2,000 original photos, are used. It has very high accuracy over 90%. (On 1 million training dataset images, 96% sensitivity and 90% specificity.) Further two tests with photos not used in training showed over 97% detection rates. Notice that this is not 100% detection rate! So care should be taken. Accuracy of detecting not-cancer images as not-cancer is high. But less accurate than detecting cancer images. Note that some bad skin infection photos look alike cancer on visual level. So this system may rate as high risk. When conditions don't improve, always check with a doctor!
Skin Cancer Types Detected
Three main types of skin cancer are detected: melanoma, basal cell carcinoma (BCC) and squamous cell carcinoma (SCC);
(1) Melanoma is the most dangerous skin cancer. It can grow quickly and spread. Melanoma can become life-threatening in as little as six weeks. When detected or in doubt, see a doctor immediately!
(2) Basal cell carcinoma (BCC) is the most common but least dangerous. It grows slowly on head, neck and upper body.
(3) Squamous cell carcinoma (SCC) can grow over months and spread to other parts of body.
About Data and Privacy