A

Project Report

On
Skin Disease Detection Application

“SKINCURE”

Submitted by:

Prajjaval Singh (21636)

Rohil Pandey (21645)
Amit Joshi (21612)
Shreya Singhal (21656)

Submitted to

Prof. D. L. Gupta
Prof. Sohit Shukla

For Partial Fulfilment of Award of Degree

Batchelor of Technology

Department of Computer Science & Engineering

Kamla Nehru Institute of Technology, Sultanpur

Affiliated to Dr. A.P.J. Abdul Kalam Technical University, Lucknow

May, 2025
 DECLARATION

We hereby declare that this submission is our own work and that, to the best of our
knowledge and belief, it contains no material previously published or written by another
person or material which to a substantial extent has been accepted for the award of any other
degree or diploma of the University or other institute of higher education, except where due
acknowledgement has been made in the text.

Prajjaval Singh (21636)

Rohil Pandey (21645)

Shreya Singhal (21656)

Amit Joshi (21612)

II
 CERTIFICATE

It is to Certify that Prajjaval Singh (Roll No. 21636), Rohil Pandey (Roll No. 21645) , Shreya
Singhal (Roll No.21656) and Amit Joshi (Roll No. 21612) have carried out the project work
presented in this project report entitled “Skin Disease Detection Application as SkinCure”
for the award of Bachelor of Technology in Information Technology from Department of
Computer Science and Engineering, Kamla Nehru Institute of Technology, Sultanpur under
my guidance. The project report embodies results of original work, and studies are carried
out by the students themselves and the contents of the project report do not form the basis for
the award of any other degree to the candidates or to anybody else from this or any other
University/Institution.

Supervisor Signature Supervisor Signature

(Prof. Sohit Shukla) (Prof. D. L. Gupta)

HOD Signature
(Dr. Arvind Kumar Tiwari)

III
 ACKNOWLEDGEMENT

After bowing our heads in front of almighty god, we take this opportunity to thank all the
people who have continuously worked for the betterment and successful completion of this
project. We would like to express our gratitude in the following order:

First and foremost, we would like to pay sincere thanks to the Director of Kamla Nehru
Institute of Technology for providing us with a conducive academic environment and for
enabling us to pursue our academic aspirations.

We extend our gratitude to Dr. Arwind Kumar Tiwari, our Head of Department, Computer
Science and Engineering, for his support and guidance throughout the project.

We would like to express our deepest gratitude to Prof. D.L. Gupta and Prof. Sohit Shukla.

Department of Computer Science and Engineering, Kamla Nehru Institute of Technology,

Sultanpur, for their constant motivation and guidance. Without their support and
encouragement, this project would have been a distant dream.

Furthermore, we would like to thank the entire staff of the department for their continuous
support, guidance, and suggestions.

Lastly, we would like to thank our parents, whose unwavering love and support have been
a constant source of strength for us. We are forever grateful for their encouragement and
guidance.

IV
 ABSTRACT

SkinCure is a mobile application designed to help users detect potential skin diseases using
Machine Learning and the convenience of Android smartphones. The platform’s primary
goal is to make early detection of skin conditions accessible and easy for everyone,
especially in areas where access to dermatologists may be limited.

One of the key features of SkinCure is its simplicity. Users can take a photo of a skin issue
using their phone camera, and the app uses a trained ML model to analyze the image and
provide a preliminary diagnosis. This direct, user-friendly approach allows individuals to
get instant insights into their skin health without needing a medical appointment right away.

SkinCure aims to improve healthcare accessibility by enabling quicker self-assessments.

While it does not replace professional medical advice, it acts as a helpful first step for users
to understand their symptoms and decide if they need to seek further medical attention. By
providing an easy-to-use interface and fast analysis, the app helps bridge the gap between
patients and healthcare providers.

This project is all about leveraging technology to enhance personal health awareness and
promote timely action through smart and efficient tools.

V
 TABLE OF CONTENTS

DECLARATION

II

CERTIFICATE III

ACKNOWLEDGEMENT IV

ABSTRACT V
1. INTRODUCTION 1-2

1.1 Motivation ...................................................................................................................1

1.2 Problem Statement.........................................................................................................1
1.3 Objectives .................................................................................................................. 2
Summary ................................................................................................................... 2

2. LITERATURE SURVEY 4-5

2.1 Methodologies .............................................................................................................. 4
Summary .................................................................................................................. 5

3. SYSTEM REQUIREMNETS 6-7

3.1 Introduction .................................................................................................................6
3.2 Software / Hardware Requirement...............................................................................6
Summary ...................................................................................................................7

4. SYSTEM DESIGN 8-11

4.1 Introduction ...................................................................................................................8
4.2 Proposed System ...........................................................................................................8
4.3 Data Flow Diagram .......................................................................................................9
Summary ................................................................................................................. 11

5. PHASES OF DEVELOPMENT 12-13

VI
6. FLOW CHART 14-17
7. IMPLEMENTATION 18-21
7.1 Introduction ..............................................................................................................18
7.2 System Design .............................................................................................................18
7.3 Algorithm ..............................................................................................................18

7.4 Architectural Components ...........................................................................................18

7.5 FEATURE Extraction ..................................................................................................18
7.6 Packages/ Libraries Used .............................................................................................19
Summary .............................................................................................................21

8. SYSTEM TESTING 22-23

8.1 Introduction ..............................................................................................................22

8.2 Test Cases ..............................................................................................................22

8.3 Result 23

8.4 Performance Evaluation .................................................................................................... 23

Summary ............................................................................................................................23

9. CONCLUSION 24
10. FUTURE SCOPE 25
REFERENCES 26

VII
 Table of Figures

Figure no. Name of Figures Page no.

1 O level Dataflow Diagram 9

2 1 level Dataflow Diagram 10

3 Flowchart Diagram 14

4 Process flow Diagram 15

5 Use Case Diagram 16

6 ER Diagram 17

7 Logo 19

8 Camera and CPU interface 20

9 Displaying Result 21

VIII
 INTRODUCTION

In an era defined by digital innovation and the growing role of artificial intelligence in
healthcare, traditional approaches to medical diagnosis are undergoing a significant
transformation. SkinCure emerges as a pioneering solution, poised to revolutionize how
individuals monitor and address their skin health through mobile technology and intelligent
automation.
SkinCure represents more than just a mobile app; it embodies a vision of accessible,
proactive healthcare for all. At its core, SkinCure seeks to overcome the challenges
associated with early detection of skin diseases—such as limited access to dermatologists,
long wait times, and lack of awareness—by providing a seamless and intelligent diagnostic
platform. Through the integration of Android development tools and machine learning
algorithms, SkinCure breaks down conventional barriers to healthcare by enabling users to
conduct initial skin assessments directly from their smartphones.
This digital health ecosystem empowers users to capture images of skin anomalies and
receive fast, AI-driven diagnostic insights based on trained ML models. By offering a
straightforward interface and instant feedback, the app enhances early detection and
encourages timely medical consultation when necessary
In this introduction, we explore the fundamental principles and transformative features of
SkinCure—demonstrating how it represents a paradigm shift in personal health
management. From the intuitive image-capture and analysis process to the intelligent,
privacy-conscious feedback system, SkinCure embodies a commitment to accessibility,
efficiency, and user empowerment in the field of digital health.

1.1 Motivation
The motivation behind the development of Shop It emerged from the growing demand for
a seamless and efficient online shopping experience that caters to both customers and
sellers. With the rapid digitalization of commerce, many users still face difficulties such as
limited product choices, slow website performance, and poor mobile usability. Shop It aims
to address these issues by creating a modern, user-centric e-commerce platform that
simplifies the buying and selling process, enhances accessibility, and offers a personalized
and secure shopping environment.
1
1.2 Problem Statement
Traditional methods of diagnosing skin diseases often involve inefficiencies such as
delayed appointments, limited access to dermatologists, and a lack of immediate feedback
for patients. These barriers can result in the progression of untreated conditions and
increased healthcare burdens. SkinCure addresses these challenges by offering a
centralized, intelligent mobile platform that enables users to perform preliminary skin
assessments using their smartphones.
By leveraging machine learning technology, SkinCure allows users to analyze skin
anomalies in real-time, providing instant feedback and suggested condition types. This
empowers individuals to take early action, reduces unnecessary delays in diagnosis, and
supports better-informed decisions about when to seek professional medical care. With its
intuitive Android interface, SkinCure makes skin health monitoring more accessible,
efficient, and proactive for everyone.

1.3 Objectives
The key objectives of Shop It are as follows:
• To develop a user-friendly e-commerce platform that ensures smooth navigation and
interaction for both buyers and sellers.
• To provide a wide range of products organized by categories, brands, and user
preferences.
• To implement efficient search and filtering mechanisms to help users find desired
products quickly.
• To support secure user authentication, order placement, and payment processing.
• To enhance user trust and transparency through product reviews, seller ratings, and
return policies.
• To enable real-time inventory management and sales tracking for sellers.
• To ensure responsive design for consistent performance across all devices and
screen sizes.

2
Summary
SkinCure is a Skin Disease Detection Application designed to transform how individuals
monitor and manage their skin health. By harnessing the power of machine learning and
Android-based mobile development, SkinCure aims to simplify the process of identifying
potential skin conditions, making early detection more accessible and efficient for users.

Utilizing technologies such as TensorFlow (for ML model integration) and Android SDK
tools, SkinCure delivers real-time diagnostic insights through an intuitive, user-friendly
interface. With its intelligent image analysis capabilities and responsive design, the
application enhances the speed, accuracy, and convenience of skin assessments—Through
its smart features and focus on accessibility, SkinCure strives to improve awareness,
promote early intervention, and bridge the gap between individuals and professional
dermatological care.

3
 LITERATURE SURVEY

2.1Methodologies

In the development of SkinCure, several methodologies and approaches were employed to

ensure the application's accuracy, usability, and accessibility. These methodologies
include:

6.1 User-Centered Design (UCD): SkinCure was developed with a strong emphasis on the
needs and behaviors of end users—individuals seeking quick and reliable skin health
insights. Through iterative design cycles and continuous user feedback, the interface and
features were refined to ensure ease of use, clarity of results, and overall user satisfaction.

6.2 Agile Development: Agile methodology guided the development process of SkinCure,
allowing for adaptive planning, evolutionary development, and continuous improvement.
With regular sprints and testing phases, new features were implemented quickly, and user
feedback was integrated promptly to enhance app functionality and performance.

6.3 Machine Learning Integration: A pre-trained machine learning model was integrated to
perform image classification of skin conditions. Technologies such as TensorFlow Lite
were used for efficient on-device processing. This integration ensures fast, real-time
diagnosis while maintaining user data privacy and reducing dependency on internet.

6.4 Android Native Development: SkinCure was built using native Android development
tools such as Android Studio and Kotlin/Java to ensure optimal performance,
responsiveness, and compatibility with a wide range of Android devices. This approach
provided full access to device capabilities like the camera for capturing skin images and
local storage for saving user result.

6.5 Responsive and Accessible UI Design: To accommodate users with different levels of
technical familiarity, the app was designed with a clean, intuitive, and accessible interface.
The layout, typography, and navigation were optimized for various screen sizes and
resolutions, ensuring a smooth experience on smartphones and tablets alike.

4
Summary
The literature survey conducted for the development of SkinCure emphasized the
significance of user-centered design, agile development practices, intelligent machine
learning integration, and efficient mobile application architecture. Research underscored
the growing role of artificial intelligence in healthcare, particularly in early disease
detection, and the need for mobile solutions that are both accessible and reliable.

By incorporating these methodologies into the development process, SkinCure aims to

deliver a robust, user-friendly mobile application that effectively addresses the challenges
associated with early skin disease detection. The survey also highlighted the importance of
on-device processing for privacy and speed, the necessity of intuitive user interfaces in
healthcare apps, and the growing demand for tools that bridge the gap between patients and
timely medical consultation. Through these insights, SkinCure is designed to empower
users with intelligent, real-time dermatological support directly from their smartphones.

5
 SYSTEM REQUIREMENTS

3.1 Introduction

System requirements define the necessary hardware and software components needed for
the successful development, deployment, and operation of SkinCure, a mobile-based skin
disease detection application. This section outlines the essential specifications required for
both the development environment and end-user usage.

3.2 Software and Hardware Requirements

3.2.1 Software Requirements:

• Android Studio: The official integrated development environment (IDE) for Android
development. It supports features such as real-time code analysis, emulator testing, and
Gradle-based build support.

• Programming Languages:
o Java/Kotlin: The primary programming languages used for building native Android
applications.

• TensorFlow Lite:
o A lightweight version of TensorFlow designed for mobile and embedded devices.
Used to run on-device machine learning models for real-time image classification
and skin disease detection.

• Firebase:
o Realtime Database: Stores and synchronizes user data in real time.
o Authentication: Manages secure login through email, phone number, or third-
party providers.
o Cloud Storage: Stores images uploaded by users for analysis.
o Crashlytics: Provides real-time crash reports for improving app stability.
o Cloud Messaging: Sends push notifications for results, updates, or reminders.
o Analytics: Tracks user interaction for performance evaluation and future
improvements.

6
• OpenCV (Optional):
o Used for preprocessing images (e.g., noise reduction, resizing, and contrast
enhancement) before passing them to the ML model.

• Internet Connection:
o Required for Firebase access, initial model download (if not bundled with the
app), and app updates.

• GitHub:
o Version control and collaboration platform used for managing source code, tracking
issues, and maintaining project history through repositories, commits, branches,
and pull request.

3.2.2 Hardware Requirements:

• Android Smartphone or Tablet:

o Minimum OS: Android 7.0 (Nougat) or higher
o Recommended RAM: 2 GB or more
o Storage: At least 200 MB of free space for app installation and data caching
o Camera: Functional rear-facing camera with minimum 5MP resolution for capturing
clear skin images

• Development Machine (for developers):

o Processor: Intel i5 or equivalent
o RAM: Minimum 8 GB
o Storage: At least 10 GB free for Android Studio, SDKs, emulators, and project files
o Operating System: Windows, macOS, or Linux with Android Studio installed

Summary
The system requirements for SkinCure ensure smooth development and efficient user
experience on Android devices. By leveraging native mobile technologies, machine
learning frameworks like TensorFlow Lite, and cloud services through Firebase, the app
provides a scalable and intelligent solution for early skin disease detection. Compatibility
with a wide range of Android devices further enhances the app's accessibility and usability
for diverse users.

7
 SYSTEM DESIGN

4.1 Introduction
System design involves the architectural planning and structuring of SkinCure, a mobile
application intended to assist users in identifying skin diseases using machine learning.
This section presents the design approach, system components, and interaction flow that
guide the development and operation of the application, ensuring it is efficient, user-
friendly, and technically robust.

4.2 Proposed System

The proposed system for SkinCure follows a mobile-centric architecture, integrating
on-device machine learning, cloud services, and user interface components for
seamless operation and accurate skin disease detection.

Mobile Frontend (Android App)

• Developed using Kotlin or Java in Android Studio.
• User Interface (UI) is designed with Material Design principles for a clean and
intuitive experience.
• Enables users to:
o Receive predictions from the ML model.
o View suggestions or further action steps.
o Access their profile and history of scans.
o Capture or upload skin images.

Machine Learning Module

• Uses TensorFlow Lite for on-device inference.
• A pre-trained Convolutional Neural Network (CNN) is integrated into the app to
classify skin images into predefined disease categories (e.g., eczema, psoriasis, acne).
• Image preprocessing (resizing, normalization) is performed before predictions.

Offline and Real-time Capability

• The app supports offline functionality by caching recent scans and model data.
• Firebase ensures real-time synchronization when the user regains internet access.

8
Responsive and Device-Compatible Design
• Designed for a wide range of Android devices (phones and tablets) with various screen
sizes.
• UI components adapt dynamically to ensure accessibility and usability regardless of device
resolution.

9
4.3 DATA FLOW DIAGRAM

4.3.1 O level Data Flow Diagram

Figure 1: O level Dataflow diagram

SkinCure is a mobile application that helps users detect skin diseases using machine
learning. Users authenticate through the app, upload images of their skin condition, and
receive a diagnosis based on machine learning analysis. The app uses TensorFlow Lite for
real-time image processing, classifying skin conditions like eczema or acne. Users can view
their results, track changes, and access past diagnoses. Optionally, data is stored in Firebase
for history tracking, and notifications are sent for follow-ups. The system offers a user-
friendly interface for quick and accessible skin condition detection.

10
4.3.2 1-Level Data Flow Diagram

The SkinCure system is designed to provide users with an easy-to-use platform for
detecting skin diseases through machine learning on Android. The system’s key
components include:

Figure 1: 1 level Dataflow Diagram

Admin: This entity includes administrators who are responsible for managing products,
handling orders, creating or applying coupon codes, and monitoring stock levels to ensure
smooth operation of the platform.

Product Classification: This process involves organizing products based on categories,

price range, and availability, helping users efficiently search and filter products according
to their preferences.

11
Order Management System: The core system that oversees product management, order
tracking, payment processing, and customer reviews. It ensures smooth operations by
maintaining an up-to-date database of products, user orders, and reviews in MongoDB, and
it handles secure transactions through the integrated payment gateway.

Summary
The system design of SkinCure focuses on offering users an efficient and responsive
platform for skin disease detection using machine learning on Android. By leveraging
Firebase for data storage and real-time updates, and TensorFlow Lite for on-device image
classification, SkinCure ensures accuracy, security, and ease of use. The architecture
supports seamless interaction between users and the app, delivering a user-friendly
experience with quick results .

12
 PHASES OF DEVELOPMENT
The Software Development Life Cycle (SDLC) for the SkinCure skin disease detection
application involves several crucial phases to ensure a seamless and effective development
process:

1. Planning: The initial phase focuses on defining the project's scope, objectives, and
requirements. This includes identifying user needs for skin disease detection, market
research to understand existing solutions, and creating a roadmap for development. The
planning phase also involves selecting appropriate machine learning models and
frameworks for efficient skin condition detection.

2. Analysis: In this phase, detailed requirements are gathered and analyzed. The team
evaluates the target user base (patients, dermatologists, etc.), identifies critical features
(image upload, real-time diagnosis, etc.), and documents the technical specifications
required for developing the Android app. This phase also includes determining the
type of skin conditions to be detected and the data requirements for the machine
learning model.

3. Design: The design phase focuses on creating the architecture of the application and
machine learning model. The design includes developing wireframes for the user
interface (UI/UX), defining the user journey, designing database structures for storing
medical images and results, and specifying the integration of TensorFlow Lite for
machine learning. This phase also includes designing a responsive app layout for
optimal usage on various Android devices.

4. Development: This phase involves the actual coding and implementation of the
application. Android developers write the necessary code to handle user inputs (image
uploads), interface with the machine learning model (TensorFlow Lite), and display
results. Developers also integrate Firebase for secure storage of user data and past
diagnoses. The model is trained and integrated into the app, ensuring efficient real-
time classification of uploaded images.

13
5. Testing: Rigorous testing ensures the skin disease detection application operates
effectively and meets user needs. This includes unit testing to verify individual
components, such as the machine learning model and image upload functionality,
integration testing to ensure seamless interaction between the app’s modules, system
testing to evaluate the overall functionality and performance of the complete system,
and user acceptance testing (UAT) to confirm that the app satisfies real user
requirements and performs well in real-world conditions.

6. Deployment: Once the app passes testing, it is deployed to the Google Play Store
for public use. This phase includes configuring the production environment, ensuring
that Firebase and other services are set up correctly, and making the application
available for download by users. Additionally, server-side configurations are
optimized to ensure smooth operation.

7. Maintenance: After the app is live, continuous maintenance and support are
necessary to address any issues that arise, including bug fixes, performance
improvements, and the addition of new features. This phase also involves updating the
machine learning model with new data, enhancing the detection capabilities, and
refining the app based on user feedback to improve its accuracy and usability.

14
 FLOW CHART

6.1 Flow Chart Diagram

Figure 3: Flow chart

The flow chart for the skin disease detection application begins with the start point, where
users are prompted to either log in if they already have an account or register by providing
their details. Once logged in, users proceed to upload an image of the affected skin area
using their mobile device. The uploaded image is then processed by a machine learning
model, which analyzes the image to detect possible skin conditions. . The system displays
the predicted result along with basic information about the condition, including
symptoms and suggested next steps. Users can then choose to consult a dermatologist
through integrated contact options for further advice or treatment.

15
6.2 Process Flow Diagram

Figure 4: Process Flow Diagram

The process flow for the Skin Disease Detection Application begins at the start point,
leading to the home screen where users are given the option to register or log in. Users
may log in as a registered user or continue as a guest. Registered users must provide their
email ID and password to access the app. Once logged in, users are directed to the main
dashboard, where they can access key features such as uploading skin images for analysis,
viewing diagnosis history, or updating profile settings. Upon uploading an image, the app
processes it using a machine learning model to detect potential skin conditions and
presents the results along with relevant details. This streamlined flow ensures a user-
friendly experience for quick and effective skin health assessments.

16
6.3 Use Case Diagram

Figure 5: Use Case Diagram

The process flow for the Skin Disease Detection Application begins at the start point,
leading to the home screen where users are given the option to register or log in. Users
may log in as a registered user or continue as a guest. Registered users must provide their
email ID and password to access the app. Once logged in, users are directed to the main
dashboard, where they can access key features such as uploading skin images for analysis,
viewing diagnosis history, or updating profile settings. Upon uploading an image, the app
processes it using a machine learning model to detect potential skin conditions and
presents the results along with relevant details. This streamlined flow ensures a user-
friendly experience for quick and effective skin health assessments.

17
6.4 ER Diagram

Figure 6: ER Diagram

The ER (Entity-Relationship) Diagram for the Skin Disease Detection Application

represents the relationships between key entities such as User (Patient), Admin,
Dermatologist, and Diagnosis Report. Users register and upload skin images, which are
analyzed by the system using a trained machine learning model. The Diagnosis Report
entity stores the results, including predicted skin conditions and confidence scores. Admins
oversee the system, manage user data, and maintain the ML model database, while
dermatologists can access reports for review and consultation. This structured relationship
enables efficient data flow, secure access.

18
 IMPLEMENTATION

7.1 Introduction
The implementation phase of the Skin Disease Detection Application involves
transforming the planned system design into a fully functional mobile application. This
includes integrating machine learning models, user interfaces, backend infrastructure,
and mobile-specific components. The application is built for Android using Android
Studio, with Python for model development and Firebase for backend services..

7.2 System Design

Based on the proposed architecture, the Android front-end was developed using XML
and Java/Kotlin for user interaction, while the backend includes Firebase for user data
storage and authentication. The ML model, trained in Python, is integrated into the app
using TensorFlow Lite for on-device inference.

7.3 Algorithm
The core algorithm used is a Convolutional Neural Network (CNN) trained on labeled
skin disease images. This model classifies input images into various skin conditions
with confidence scores. Preprocessing algorithms are used for image resizing,
normalization, and augmentation to improve prediction accuracy.

7.4 Architectural Components

The main components of the skin disease detection application include the Android-
based frontend for capturing or uploading images, displaying results, and user
navigation; the embedded TensorFlow Lite (TFLite) machine learning model for
offline image classification; a Firebase backend that manages user authentication,
optional image storage, and report saving; and the controller logic that ensures
seamless data flow and coordination between the user interface, machine learning
model, and backend services.

7.5 Feature Extraction

Key features developed include:
• Image capture or gallery upload.
• On-device ML prediction and display of disease classification.

19
7.6 Packages/Libraries Used
TensorFlow & TensorFlow Lite: For model training and Android integration.

Android Jetpack Libraries: For lifecycle management and navigation.

CameraX: For efficient image loading and display.

7.7 SCREENSHORT
7.7.1 Logo of Application

Figure 7: Logo

The logo's design embodies a clear vision, mirroring how our app helps users see and
understand their skin conditions. It has been designed to visually represent the core purpose
and functionality of the project-health, technology, and accessibility. It features a clean and
minimalistic design.

20
7.7.2 Preview of Camera and CPU interface

Figure 8: Camera and CPU interface

The image processing module is a crucial part of the skin disease detection system. It
prepares the input images for accurate analysis by the machine learning model. The screens
below demonstrates the various stages involved in the image processing workflow within
the application.The user captures a photo or uploads an existing image of the affected skin
area using the app interface.The image undergoes several preprocessing steps, such as
resizing, normalization, and noise reduction. This ensures uniform input dimensions and
quality for the model.In some cases, segmentation techniques are applied to isolate the
region of interest (i.e., the skin lesion) from the background to improve classification
accuracy.The processed image is then analyzed to extract key features such as texture, color
patterns, and shapes that are relevant to skin disease classification.

21
7.7.3 Displaying Result

Figure 9.Displaying Result

The result display screen is the final and most user-focused stage of the skin disease
detection process. After an image is captured and processed by the application, the machine
learning model analyzes it and predicts the most likely skin condition. The result is then
presented to the user in a clear and informative manner. It predicts the disease name of
certain type allowing the user to quickly understand the diagnosis. The processed image is
shown alongside the result for visual confirmation and context.

Summary
The implementation of the Skin Disease Detection App involves integrating Android
development with machine learning and cloud technologies. Through well-defined
architecture and the use of essential libraries, the application effectively provides real-time,
offline skin disease predictions and a seamless user experience.

22
 SYSTEM TESTING

8.1Introduction

System testing is a critical phase in the development of the Skin Disease Detection App,
ensuring that all components—from machine learning inference to user interface—function
as intended. This section highlights the testing strategy, including defined test cases, their
outcomes, performance evaluation.

8.2 Test Cases

Test cases were designed to verify core functionalities such as image upload, model
prediction, UI responsiveness, and Firebase integration. Scenarios include user
authentication, image processing, result display, and data storage.

Example Test Cases:

1. Image Prediction
• TC_03: Upload Skin Image
o Steps: Open app → Select/Take Image → Submit.
o Expected: Image is passed to TFLite model, and result is displayed.
• TC_04: Invalid Image Upload
o Steps: Upload blurred or unrelated image.
o Expected: Error message prompting user to upload a valid skin image.

2. Data Handling and Reporting

• TC_05: Save Report to Firebase
o Steps: Complete diagnosis → Save report.
o Expected: Report data is saved and retrievable.

3. Performance Test
• TC_06: Model Inference Time
o Steps: Upload image and record prediction time.
o Expected: Prediction completes under 2 seconds on average.

23
8.3 Result

Results showed successful execution of major test cases, including real-time diagnosis and
Firebase integration. Minor UI bugs were fixed, and performance was optimized.
Example Result:
• Test Case: Image Prediction — Result: Pass
• Comments: Model returns accurate prediction in 1.5 seconds.

8.4 Performance Evaluation

Stress and load testing confirmed the app’s ability to handle concurrent image processing
and data retrieval efficiently. The app performed reliably across multiple Android devices
with varied hardware specifications.

Summary
System testing ensured the Skin Disease Detection App meets its objectives in terms of
usability, performance, and reliability. By addressing bugs early and validating critical
paths, the application is prepared for real-world deployment with minimal risk.

24
 CONCLUSION
The development of the Skin Disease Detection Application using Machine Learning and
Android marks a significant contribution toward the integration of technology in
healthcare, especially in dermatology. This project aimed to provide a convenient, cost-
effective, and accessible tool that enables users to perform preliminary analysis of skin
conditions directly from their smartphones. By leveraging the power of machine learning
for image classification, the application assists in the early detection of various skin
diseases, potentially leading to quicker diagnosis and timely treatment.
The machine learning model was trained using a diverse dataset of labeled skin disease
images. Techniques such as image preprocessing, feature extraction, and model
optimization were employed to improve accuracy and reliability. The integration with the
Android platform ensures that the solution is portable, user-friendly, and widely accessible
to a broad population, including those in remote and underserved areas where access to
dermatologists may be limited.
Through extensive testing and validation, the application demonstrated satisfactory
performance in identifying common skin conditions. The user interface was designed with
simplicity in mind, making it easy for users with little to no technical background to
navigate the app and understand the results.
Despite its success, the project also revealed areas for improvement. The accuracy of the
predictions is still dependent on the quality and variety of the training data. Future
enhancements may include expanding the dataset, incorporating more complex deep
learning models such as convolutional neural networks (CNNs), and enabling real-time
consultation with medical professionals. Moreover, ensuring the app complies with
medical data privacy regulations and ethical standards will be crucial as it scales.

25
 FUTURE SCOPE
The future scope of SkinCure is expansive and offers numerous opportunities to further
enhance its functionality, user engagement, and market expansion:
• Advanced ML Models: Incorporate deeper CNN architectures or transformer-based models for
improved diagnostic accuracy.
• Telemedicine Integration: Enable real-time consultation with dermatologists via video or chat.

• Multi-Disease Classification: Expand model capabilities to detect a broader range of skin

conditions.
• Offline Diagnosis Enhancements: Improve TFLite model for enhanced offline prediction.

• Geotagging and Reporting: Enable anonymized location-based data aggregation for public
health monitoring.
• iOS Version: Extend support to iOS users.

• Health Record Integration: Sync diagnosis results with EHR systems.

• Security and Privacy Improvements: Implement encryption for stored images and user reports.

26
 REFERENCES

Here are the direct URLs for your references:

1. TensorFlow Lite Documentation. [Online]. Available: https://www.tensorflow.org/lite
2. CameraX API Documentation. [Online]. Available:
https://developer.android.com/training/camerax

3. Harvard Dataset link:

https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7 910/DVN/DBW86T
4. "Deep Learning for Dermatology: A Review," Journal of Medical Internet Research,
2022.
5. "Real-time Skin Disease Detection Using Mobile Phones," International Journal of
Computer Vision, 2021.
6. https://www.w3.org/
7. https://developer.android.com/

27