Real time based Melanoma detection by using Non invasive technique through low cost dermatoscope
Our skin is the largest of our organs and plays a vital role in maintaining our lives. Like any other organ of the body, the skin is also subjected to cancer in that Melanomas are considered as the deadliest form of skin cancer because of its ability to spread to other organs more rapidly its early
2025-06-28 16:34:42 - Adil Khan
Real time based Melanoma detection by using Non invasive technique through low cost dermatoscope
Project Area of Specialization Artificial IntelligenceProject SummaryOur skin is the largest of our organs and plays a vital role in maintaining our lives. Like any other organ of the body, the skin is also subjected to cancer in that Melanomas are considered as the deadliest form of skin cancer because of its ability to spread to other organs more rapidly its early detection and intervention implicates higher chances of curing the disease. Melanomas are asymmetrical and have irregular borders, notched edges, and color variations, so analyzing the shape, color, and texture of the skin lesion is important for melanoma early detection and prevention. Notwithstanding the reliable diagnosis of skilled and experienced medical practitioners, their diagnosis takes lots of time, efforts and costly biopsy. Due to overlapping and heavy variations of present artifacts( like hair, skin lines, blood vessels, etc) in benign and malignant lesions, real-world professionals cannot always provide a reliable diagnosis First, the advent of dermoscopy has enabled a dramatic boost in the clinical diagnostic ability to the point that melanoma can be detected in the clinic at the very earliest stages. Furthermore, the development of advanced technologies in the areas of image processing and machine learning has given us the ability to allow distinction of malignant melanoma from the many benign mimics that require no biopsy however such devices are too costly and require experienced professionals for its usage. To make this diagnosis cost-effective and readily available every-where our proposed project idea is to design and fabricate inexpensive portable device that captures real-time image through raspberry pi with the interfacing of pi camera, illuminator used for the enhancement of images and a 12x magnifier to get magnified clear image for better observation after that deployed machine learning-based model in raspberry pi that will predict the chances of melanoma from lesion skin. The model will not only be trained on the ISBI 2017 dataset that contains almost 23906 of quality-controlled dermoscopic images of skin lesions having different attributes but also dataset obtained through low-cost dermotoscope that contain different non-cancerous lesion. To maintain the patient record, the captured images will be sent to google cloud for storage purposes.
Fig 01: Project Flow Diagram
Project Objectives
The key project objectives are stated here:
- To design and fabricate a prototype for skin cancer detection that captures magnified image of skin lesion
- To implement image processing and Machine Learning techniques to make a system that can recognize skin cancer from lesion
- To help patients to prevent melanoma in early stages.
Our project is mainly based on image processing and machine learning for that real-time image will be captured through low-cost dermatoscope. To make classification model there are several steps to be followed as:
- The captured raw image will be pre-processed to improve the quality of the image by removing unwanted artifacts and image enhancement
- The pixels of image are segmented out as cancerous pixels or normal skin
- The features (like Asymmetry, Border Irregularity, Color Variation, Diameter of suspicious lesion) will be extracted from that segmented image for better classification
Once the model is trained now the interfaced pi camera will capture the patient’s suspected lesion image, this image will be called in a trained machine learning model for classification and will also be sent to google cloud as well for patient’s record purpose. The implementation methodology is depicted in Figure.2. We have consulted a doctor at LUMS Jamshoro, who will assist us in acquiring real-time images of skin related complications and also provide a medical assessment on the algorithm results.
Fig 02: Implementation Method
Benefits of the Project
The healthcare sector has long been an early adopter of and benefited greatly from technological advances. These days, machine learning (a subset of artificial intelligence) plays a key role in many health-related realms not only the development of new medical procedures but also in the enhancement of disease diagnosis with more accuracy in less time. Following are the benefits of this project:
- Doctors: This project will help doctors to diagnosis more effectively, which ultimately leads to better outcomes, costs lower than care, and increase patient satisfaction.
- Skin cancer Patients: Even a normal mole can be cancerous but today people don’t know about their skin they only believe what they see but in actual it might be something else so this project will provide a safe way to skin cancer patients for earlier diagnose.
The final product will be low-cost dermotoscope comprises of raspberry pi module, pi camera, magnifier, illuminator and display to provide ease in real time-based melanoma detection. In this raspberry pi works as a microprocessor with an interfaced pi camera for capturing lesion images with 12x magnifier to magnify the image finally the illuminator will be used to enhance the image effect then this image will be loaded in machine learning model through Graphical User Interface application to classify the image as melanoma or non-melanoma cancerous type. The model will be trained on Quadro 600 GPU for quick classification then captured images will also send to google cloud for storing, which will be used as patient’s record purpose.
Fig 3: Proposed GUI
Final Deliverable of the Project Hardware SystemCore Industry HealthOther Industries IT , Medical , Health Core Technology Artificial Intelligence(AI)Other Technologies Artificial Intelligence(AI)Sustainable Development Goals Good Health and Well-Being for PeopleRequired Resources
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 79918 | |||
| Respberry Pi 4 Borad | Equipment | 1 | 7500 | 7500 |
| Respberry Pi Camera | Equipment | 2 | 5000 | 10000 |
| Carson MicroFlip 100x-250x LED and UV Lighted Pocket Microscope | Equipment | 1 | 9558 | 9558 |
| Quadro 600 GPU | Equipment | 1 | 30000 | 30000 |
| SD Card (32GB) | Equipment | 3 | 900 | 2700 |
| HDMI to micro HDMI convertor | Equipment | 3 | 220 | 660 |
| Charger (port c) | Equipment | 1 | 500 | 500 |
| Cardboard with wheels attachment | Equipment | 1 | 3000 | 3000 |
| Travelling, consultation fees etc | Miscellaneous | 10 | 1000 | 10000 |
| Phone Camera Lens,ARORY Pro Photography Lens Kit | Equipment | 2 | 3000 | 6000 |