Development of a Medical Analysis Recognition System to Enhance Clinic Efficiency
Applied computer vision and machine learning to automate the processing of medical documents
- Client
- MedDoc
- Year
- 2023
- Services
- Computer Vision
- Platform
- Python, YOLO
About project
The company 32desk approached us with the goal of developing a system for recognizing medical analyses from photographs and subsequently conducting analytics based on the obtained data. This solution aimed to automate the processing of medical documents, reduce data entry time, and decrease the likelihood of errors associated with human factors.
Solution
We collected a dataset of 10,000 images of medical analyses. To expand the data volume and increase the model's robustness to various conditions, we applied data augmentation methods: rotation, distortion, adding noise, and other techniques. A team of specialists labeled the dataset using a specially developed web interface, ensuring high quality and accuracy of the training data. Based on the prepared dataset and utilizing YOLO (You Only Look Once) technology, we trained a model capable of converting images of medical analyses into a structured tabular format. This allowed for the automatic extraction of key indicators and data from analysis photographs.
Result
The pilot project was successfully implemented in a network of dental clinics across the country. The system automated the process of processing medical analyses, significantly reduced data entry time, and decreased the likelihood of errors. Clinics received a tool for quick analytics and decision-making based on up-to-date medical data, which improved patient service quality and increased the efficiency of medical staff.
