How imakashsahu/third-eye-final-year-project works
Project Type & Purpose
• This is a forensic face sketch construction and recognition application. It is designed to allow users (e.g., law enforcement personnel) to drag and drop facial components to construct a composite sketch – a process that traditionally requires a forensic artist. The system then automatically matches the composite sketch with entries from the police database using deep learning on a cloud infrastructure, notably integrating AWS S3 for storage and AWS Rekognition for processing.
• The project is a standalone application with a mix of client-side (interactive sketch creation) and backend components (deep learning recognition and cloud services).
Analysis of the File Structure
• The “Face Sketch Elements” directory and its subdirectories (eyebrows, eyes, hair, head, lips, more, mustach, and nose) represent the visual building blocks that the application uses for sketch construction.
• The existence of various PDF reports (Final Report, Technical Paper Certificates, Published Paper) show that there is also comprehensive documentation regarding the project.
• Although the file tree does not explicitly separate code into typical folders like “src” or “backend”, the folder structure clearly organizes graphical elements, supporting the UI/UX aspects integral to the application.
Using README Insights
• The README emphasizes the drag and drop interface, use of AWS cloud components (S3 for storage and Rekognition for image analysis), and deep learning for matching sketches against a police database.
• There are walkthrough videos and detailed run guides that indicate the system has both a user-facing part and a cloud-based processing backend.
a. Main Components:
• User Interface (UI): A front-end module where users interact with the system using a drag and drop interface to assemble face sketches.
• Sketch Elements Repository: A module (conceptually represented by the “Face Sketch Elements” folder structure) that contains all visual elements (eyes, eyebrows, etc.).
• Deep Learning Recognition Engine: The component responsible for processing the user-created sketch using deep learning models for facial matching.
• Cloud Infrastructure & External Services: AWS S3 (for storage), AWS Rekognition (for image analysis), and the Police Database (or any forensic data source) to match the sketches.
• Integration/Control Module: Links the UI, recognition engine, and cloud components, handling data orchestration and user requests.
b. Relationships and Interactions:
• The UI allows users to select and arrange facial components (data flow from UI to Sketch Elements Repository).
• Once the composite sketch is created, it is sent to the Recognition Engine.
• The Deep Learning module may store data temporarily in AWS S3 and then use AWS Rekognition to process and compare the image against the police database.
• Results (such as potential matches) are then returned to the user through the UI.
c. Architectural Patterns & Design Principles:
• Separation of Concerns: Clearly distinguish between the presentation layer (UI and face sketch components) and the business logic/data processing layer (deep learning engine, cloud integration).
• Modular Design: Each component – from sketch libraries to cloud services – is treated as a self-contained module that interacts through well-defined interfaces.
• Cloud Integration: Emphasize the use of AWS services as external modules, establishing a boundary between in-house modules and external dependencies.
d. Technologies and Frameworks:
• UI/UX framework (e.g., if a web app, then HTML/CSS/JavaScript frameworks might be in use).
• Deep Learning frameworks (which are referenced as part of the deep learning recognition module).
• AWS S3 and AWS Rekognition for cloud-based storage and image analysis.
• Any middleware or backend service bridging the user actions to cloud processing (the integration layer).