Challenges
Blood reports do not always follow a common structure. Laboratories may use different layouts, test labels, units, and formatting, making it difficult to collect their information within one digital system.
The client's existing process depended heavily on manually reviewing reports and entering the required values. This took time, increased the possibility of transcription mistakes, and made historical information harder to locate when users needed to compare reports.
The application needed to:
- Accept reports submitted as PDFs or images
- Process reports with different layouts
- Extract and organize relevant test information
- Clean and standardize the captured data
- Store current and historical patient records
- Present the information through an intuitive web interface
- Reduce reliance on paper-based report management
Solution
Lucent Innovation began by reviewing how information appeared across the supported blood reports. This helped the team identify the fields that needed to be captured and define a common structure for organizing test names, values, units, and other available report details.
The team developed a Python backend using FastAPI to manage file uploads and report-processing requests. OpenCV-Python was used to prepare image-based reports for extraction, including scanned files and photographs that required image processing before their content could be captured.
After the text was extracted, the processing layer cleaned the captured values and organized them into a common data structure. This allowed information from reports with different layouts to be handled more consistently within the application.
MongoDB was used to store the structured report data and available historical records. Its flexible document model made it suitable for reports that did not always contain the same tests, fields, or data arrangement.
The team then developed the frontend using ReactJS and Material UI. The interface allowed users to upload reports, review the extracted information, and access available historical records through one application.
Together, these components connected file processing, data extraction, standardization, storage, and presentation within a single workflow. The project reflects the document intelligence and workflow automation capabilities supported through Lucent Innovation's AI and ML development services.
