.png){kind=small}
OCR System
Project Type:
Date:
Project Duration:
Status:
Summary:
​
2021
5 Weeks
​
Simple UI for AI OCR analysis for an aircraft manufacturer.
AI
B2B
Done
Project Background
OCR System
OCR stands for Optical Character Recognition, and it refers to the technology that is used to convert different types of documents, such as scanned paper documents, PDFs, or images captured by a digital camera, into editable and searchable data.
​
In response to the challenges faced by a prominent player in the aviation industry, our team started the EAGLYS OCR project, strategically designed to digitize extensive technical documents and construct an intelligently searchable database.
Problem
Our customer in the aviation industry, responsible for manufacturing and maintenance of aircrafts, faced a large volume of technical and repair documents spanning various formats, including paper documents, manuals, and schematics. Their system for managing these documents was not only manual and time-consuming, but also succeptible to errors at the time of digitizing.
Solution
The bespoke solution for our client was to implement an OCR system with AI capabilities to significantly enhance document management within the aviation sector.
​
The solution involved converting paper documents and images into machine-readable text, train machine-learning models on their data to create a search algorithm, structuring and indexing the extracted text data, and building a user friendly interface to allow technicians and engineers to quickly locate specific information within the digitized documents.
My Role
As the lead designer, I was in charge of creating a simple UI aligned with client specifications, and that could seamlessly be integrated with our core AI and Secure Computing engine.
​
The OCR team included, 1 project manager, 1 AI engineer, 1 full-stack engineer, and myself as the designer.
What I Did
Vision crafting
Collaborated with the project manager to conceptualize the vision for the interface keeping in mind the customer requirements for this project and industry benchmarks.
Validation discussions
Held multiple discussions with the engineering team to validate design ideas and technical feasability of the design.
Redesign
Overhauled the initial PoC (Proof of Concept) design deployed by the engineering team to fit the client requirements.
Handoff
Delivered my designs, user flows, wireframes, and hi-fidelity prototype to the developer (full-stack engineer) to proceed with the front-end implementation.
Design Outcome
Home screen
The Home screen was redesigned using the PoC version developed by the engineering team as a basis.​​
The renewed home screen was designed as a file management system and storage application to provide familiarity to the user to manage their digital files.
​
Featuring a simple layout to visualize existing files, a multipurpose searchbar for filenames or specific characters (words, terms, diagrams, etc.), and an intuitive button to add and analyze new documents, the interface enhances user familiarity and efficiency.
File upload & AI analysis
The upload and analysis had to be smooth and the user should not worry about data pre-processing or machine learning training. To streamline the upload and analysis process, the choice was to allow the user to select a previously scanned document, choose from the pre-trained models for AI analysis, and proceed to upload the file to the system.
File view & term finder
The file view and word finder screen was also redesigned using the PoC version developed by the engineering team as a basis.
The revamped file view and word finder screen provides a familiar file management and document viewing/editing experience.
​
The top of the screen includes basi information about the file as well as a tag system to make the searches even more intuitive.
​
The main screen shows a full view of individual pages with thumbnails on the left. The right bar shows the words/terms found in the document after the AI OCR analysis, and pairing them with the active document page for better visualization. The searchbar is available to search for any word, term, or image within the document.
The Results
The feedback from the customer turned out to be very positive, as most of the users were familiar one way or another with managing digital files. The digitization of thousands of documents continues, claiming that the searching process runs smoother while needing to search for specific sections of technical documents and manuals.
My Reflection
I’m thankful for having the chance to work in this interesting project for one of the biggest aircraft manufacturers in Japan. I learned a tremendous amount during this process, especially the potential to simplify a complex process, such as AI analysis, into a simple and accesible experience for users of diverse backgrounds.
​
While a more creative solution could have been devised given more time, the emphasis on simplicity resonated with our customer's priorities, leaving me satisfied with the outcome.