Automatic visual inspection for pharmaceutics.
About the project
Our client is a Swiss startup that specializes in computer vision for the healthcare industry. They design a counting solution based on machine learning that enables automatic counting and tracking of multiple types of objects.
Radency engineer was tasked with connecting a medical labels printer into the system's workflow as well as integration of a computer vision solution into the QA process.
Computer vision integration
The computer vision integration via the gRPC protocol was done in order to perform a print quality inspection.
Label printer integration
We've integrated with the label printer and implemented the programmatic generation of the labels in Node.js so that users automatically generate adhesive labels with all necessary count information
PDF reports generator
At the end of each counting operation, the system provides detailed PDF reports with all identification data.
There is no place for mistakes in automatic visual inspection. Hence, we’ve employed a number of techniques to assure solution quality:
Covered the Node.js and Angular codebase with unit tests.
Created system test documentation
Created Swagger Open API documentation
Angular \ TypeScript \ RxJS \ Electron
Nest.js \ PostgreSQL \ gRPC \ Docker
1 software engineer on a full-time basis
After assessing candidates we handpicked a developer that completely fulfills the client’s requirements and started working immediately.
The client needed to extend their in-house development capacity with one Radency engineer. During communication, we clarified the project’s goals and requirements, defined necessary skills and desired workflow to assemble the best-fit candidate.
Integration and delivery
Radency engineer smoothly integrated into the client’s project following their preferred management approach and tools.