SMART DERMA: MOBILE APPLICATION FOR SKIN-DISEASE DETECTION USING CONVOLUTIONAL NEURAL NETWORK

Volume-09 Issue 10, October-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [114] SMART DERMA: MOBILE APPLICATION FOR SKIN-DISEASE DETECTION USING CONVOLUTIONAL NEURAL NETWORK Ronald Fernandez ORCID ID - 0009-0007-0979-6315 Professor, College of Computing Studies, Universidad de Manila, Philippines Jhon Lawrence Ami ORCID ID - 0009-0009-4158-9855 UG Students, College of Computing Studies, Universidad de Manila, Philippines Rechelle Cabaltera ORCID ID - 0009-0009-2212-8754 UG Students, College of Computing Studies, Universidad de Manila, Philippines Noel Casas ORCID ID - 0009-0007-2916-5967 UG Students, College of Computing Studies, Universidad de Manila, Philippines Jared Concengco ORCID ID - 0009-0005-1987-7034 UG Students, College of Computing Studies, Universidad de Manila, Philippines ABSTRACT Skin disease is an emergent public health problem worldwide, particularly in developing countries where healthcare access is limited. To address these challenges, the rapid advancement in Artificial Intelligence (AI) and Machine Learning (ML) technologies has introduced new opportunities to improve healthcare access and diagnostic potential. This study of Smart Derma: Mobile Application for Skin-Disease Detection Using Convolutional Neural Network (CNN) aims to address the growing need for accessible and reliable tools in dermatological self-assessment. This is a mobile application for skin disease detection using Convolutional Neural Network that will assist the users for detecting possible skin diseases through image-based analysis. It was conducted as a developmental approach at Gat Andres Bonifacio Memorial Medical Center and Justice Jose Abad Santos General Hospital and was assessed by professionals, patients, and students to evaluate its usability and reliability. However, Smart Derma mobile application is not a substitute for dermatologists but an adjunct, initial screening device to identify potential skin conditions. The study yielded both favorable and unfavorable outcomes that will help students, patients, healthcare providers and future proponents while further validation and improvement remain essential in continuing these studies. Keywords: Artificial Intelligence, Convolutional Neural Network, Digital Repository, Academic Research, Image-based Analysis INTRODUCTION Artificial Intelligence and Machine Learning have transformed many sectors, including healthcare, by offering creative solutions to intricate issues. Convolutional Neural Networks is a specialized form of Machine Learning that is particularly adept at image recognition functions by extracting patterns and features from images. Through the use of Convolutional Neural Networks, advocates and developers are able to design systems that can scan visual information, such as images of skin conditions, to offer true insights. These Artificial Intelligence-based technologies hold hopeful prospects for resolving challenges in diagnostic processes and accessibility. Volume-09 Issue 10, October-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [115] The conceived mobile-application will seek to leverage Convolutional Neural Networks to diagnose skin diseases based on user-uploaded images of skin ailments. By training the Convolutional Neural Network model on a representative and labeled set of skin conditions, the system will be able to recognize patterns linked with particular diseases. Users will be provided with an initial assessment of potential diagnoses, as well as disease history, causes and symptoms information. Although the app does not substitute professional medical consultation, it is a tool for initial assessments and facilitates decision-making in healthcare choices.. As with any Artificial Intelligence-based system, this app has strengths and weaknesses. Among its strengths include ease of access, affordability, and the capability to support users in underserved areas where there are not enough dermatologists. It also allows for rapid initial screenings, cutting down on wait times before getting medical care. Problems exist, however, including issues of bias in the training sets, accuracy of the predictions, and users mistakenly interpreting the findings. Ensuring reliability of the system and informing users of its limitations are essential to its success. The project will be of benefit to different people, such as patients, students, healthcare professionals, and advocates. Patients get access to an affordable and easy-to-use tool for initial skin disease assessments. Students pursuing Information Technology, Artificial Intelligence, or Healthcare studies can utilize the project as a reallife learning tool. Physicians would find it as a helpful additional tool to those currently in place, and activists can use it as a means to build support for more improvement in Artificial Intelligence technology use within medicine. This project fills the gap between technology and healthcare, making it easier and more efficient to manage skin conditions. OBJECTIVES This study aims to develop a Smart Derma, a mobile application for skin disease detection using Convolutional Neural Network that will assist the users for detecting possible skin diseases through image-based analysis. This study will provide the user with a diagnosis regarding their skin problem without consultation charges and long wait for appointments to consult a dermatologist. The study shall focus on the following specific objectives: 1. To assist users who self-diagnose by providing an application that generates unbiased and accurate information about skin diseases. 2. To educate users by serving as an example of a diagnostic tool that has a simple and intuitive first-time user experience. 3. To develop a diagnostic tool for detecting skin diseases and act as a preliminary assessment that is free and accessible through the user mobile phones. 4. To design a user-friendly mobile application that users would understand and easily navigate through the application to find what they’re looking for. 5. To develop a mobile application that will provide a list of possible skin diseases, the severity, and description of it. METHODOLOGY This section demonstrated the process of developing Smart Derma, a mobile application using convolutional neural network. It encompasses the planning, designing, developing, testing and launching of the results ensuring the reliability of the mobile application. Along with the systematic process the validity of using the application is gathered through consultations, interviews and surveys conducted in professionals, students and patients. The mobile application utilizes this Agile methodology for credibility, authenticity and reliability of structured repository. A. Research Design The proponents followed Agile; the agile approach that highlights the iterative development, flexibility and collaboration to adapt every change and improvement of the application. The proponents use the agile to ensure that each iteration will be an improvement. Integrating user feedback which is essential in developing a mobile application with the purpose of detecting skin diseases. Volume-09 Issue 10, October-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [116] Figure 1 AGILE of Smart Derma 1. Planning The focus during this phase is to develop a detail plan to guide the proponents such as by identifying the objective, scope of the study, also in this part is where the proponents will list down all the requirements needed in creating the application such as how the system will flow and how the proponents will collaborate with dermatologist that will help to build and verifying the results of the model. 2. Design In this phase, the proponents designed the user interface, the flow of the system ensuring that it is user friendly, and features that will be needed in the application such as the system architecture that integrates Convolutional Neural Network for analyzing images. 3. Development During this phase, the development of the application undergoes such as the development and the programming of the frontend and backend of the application. 4. Testing In testing, the system conducted various system tests to gauge the accuracy and reliability to ensure that all the features and buttons are working and it will function across different phone models, error handling mechanisms will also be made to solve issues during user interaction. 5. Deployment In deployment, the mobile application deploys to the target user group to collect the feedback, usability, and accuracy of the model's diagnosis of skin conditions. 6. Review After the deployment, the feedback that comes from the target user reviewed to know what are the users’ insights regarding the application. 7. Launch The application will be released to the users, and the proponents will provide support and will prepare for the next iteration of development of the application. RESULTS AND DISCUSSION A total of 300 responses to the survey were obtained to the outpatients in Gat Andres Bonifacio Memorial Medical Center and Justice Jose Abad Santos General Hospital whereas the students came to Universidad De Manila. The main focus of this study was to test the functionalities of the application for its accuracy and the quality of the system. A. Helping Users Feel Informed Volume-09 Issue 10, October-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [117] Figure 2 Pie Chart Showing the Application Should Help the user feel more informed According to the findings, 55.6% of the users deem it appropriate that the application should help the users to feel informed before consulting medical professionals. This indicates that users would like to know more about their condition before asking for help. B. Providing Preliminary Diagnosis Figure 3 Pie Chart Showing The application should provide a preliminary diagnosis that aligns with typical clinical assessments The findings for this question conclude that 56.3% of the users would like to trust the application’s ability to deliver a preliminary diagnosis in line with clinical expectations, this further reinforces the role of the application as a supportive tool rather than a replacement for professional care. C. Identifying Multiple Skin Diseases Figure 4 Pie Chart Showing The application should be able to identify multiple types of skin diseases, not just one or two Volume-09 Issue 10, October-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [118] With the survey, we can conclude that 57% of the users find it vital that the application should be able to identify multiple types of skin diseases, this is already a vital component of the application itself. D. Identifying Common Local Skin Diseases Figure 5 Pie Chart Showing The application should be able to identify common local skin diseases specific to our region The findings can conclude that 57.3% of the users deem it relevant that the application should have the ability to identify locally relevant skin diseases. E. Steps For A Diagnosis Is Simple Figure 6 Pie Chart Showing The steps to get a diagnosis result are simple and straightforward The results show that 62.3% of the users have highly rated that having a simple way to get a diagnosis is highly appealing for them. This shows that users value a simple and efficient process. ACKNOWLEDGEMENT We would like to express our gratitude to our Capstone Adviser, Mr. Ronald B. Fernandez, for his consistent support, productive feedback, and encouragement throughout this research. His commitment and inspiration helped us maintain our focus and aim high in our research. We would like to sincerely thank Dr. Isidro, School Clinic Doctor of the Universidad de Manila, and all the nurses and staff of the clinic for their undying support and cooperation. The implementation of this project proved very possible most especially during the system testing and data gathering. Special thanks go to our Grammarian, Mr. Josue Cotoco Miras, LPT, MAEd, for ensuring the manuscript was concise and clear. We thank Mr. Christian Oroceo Murillo, our Technical Adviser, for his valuable insights and suggestions that greatly improved our system. Volume-09 Issue 10, October-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [119] We appreciate the participation of staff and students in evaluating the system and completing our questionnaires. Their quality input and feedback have played a major role in our conclusion. And not to forget, our friends and family deserve gratitude for standing by us, supporting us, and believing in us. Only because of their support were we able to pursue this research journey. Above all, to God. CONCLUSION The feasibility analysis revealed that AI-based solutions can be incorporated into accessible healthcare solutions. This Smart Derma mobile application employs a convolutional neural network that helps the user recognize the skin diseases for initial treatment. This proved to be an efficient, user-friendly, and reliable system that is exactly what is needed in preliminary dermatological assessments in the Philippines. It is not designed to substitute dermatology experts, but is a good screening and teaching aid that allows users to make informed choices concerning their health. Minimalistic and straightforward interaction of users in the application earns a positive impact on the results of acceptance of the mobile application. The strength of Smart Derma application lies primarily in speed, resource efficiency and ease of use. Its reliance on heterogeneous datasets and its yet to be broadly clinically validated weaknesses are its core. In summary, Smart Derma has tremendous potential to create tech and healthcare bridging in underserved regions. It also cost efficient, reliable and user-friendly that more widely respondents accept the studies. Although the proponents acknowledged that this study has certain limitations, it remains open for improvements that could enhance its quality and be useful in everyday use for all users. REFERENCES [1] Véliz, C. (2022). Ethical issues of artificial intelligence in medicine and healthcare. Journal of Medical Ethics, 48(7), 371–375. https://pmc.ncbi.nlm.nih.gov/articles/PMC8826344/ [2] Liu, Y., Jain, A., Eng, C., Way, D. H., Lee, K., Bui, P., ... & Coates, M. (2020). Artificial intelligence in dermatology: A primer. Journal of the American Academy of Dermatology, 83(4), 1220–1230. https://www.sciencedirect.com/science/article/pii/S0022202X2031201X [3] Sethy, M., & Behera, H. S. (2021). Artificial intelligence in dermatology and healthcare: An overview. Indian Journal of Dermatology, Venereology and Leprology, 87(6), 709–720. https://ijdvl.com/artificial-intelligence-in-dermatology-and-healthcare-an-overview/ [4] Liu, Y., Jain, A., Eng, C., Way, D. H., Lee, K., Bui, P., ... & Coates, M. (2020). A deep learning system for differential diagnosis of skin diseases. Nature Medicine, 26(6), 900–908. https://pubmed.ncbi.nlm.nih.gov/32424212/ [5] Villarmino, M., & Medina, R. (2023). Ailbot 2.0: A symptom checker chatbot for skin diseases of young children. DLSU Research Congress Proceedings, HCT-10. https://www.dlsu.edu.ph/wpcontent/uploads/pdf/conferences/research-congress-proceedings/2023/HCT-10.pdf [6] Goel, S. (n.d.). Dermnet dataset [Data set]. Kaggle. https://www.kaggle.com/datasets/shubhamgoel27/dermnet