Europe Union

Home Case Studies Is Your PoC Too Slow? How We Optimized Algorithm Performance for Commercial Development

Is Your PoC Too Slow? How We Optimized Algorithm Performance for Commercial Development

Vizualization of improving computer vision algorithm by migration to C++ wivualized as a data flowing into a mobile phone and out of it
Calpro is a Norwegian biotech company specializing in calprotectin testing. Co-founded by the scientists who originally discovered calprotectin, the company has established itself as a leading provider of diagnostic tools centered around this biomarker. Calpro doesn’t stop innovating and making it easier for patients to test and monitor their calprotectin levels.

The company has developed a Proof of Concept (or PoC for short) for a mobile app that allows patients to test for calprotectin and monitor its levels from the comfort of their homes. Users collect a sample, scan the test strips with a phone’s camera, and receive diagnostic results directly through the mobile app.

Calpro’s next step was to make their PoC into a commercial mobile application. To accomplish that, they looked for reliable tech partners.

👋 Hey You, What Are You Working On?

Share a few details about your challenge or idea, no matter how early or complex. Our expert team will reach out to offer our expertise in terms of feasibility, budget, and scope. Let’s shape a solution for you together!

Let’s Discuss Your Project

An Innovative Self-Diagnostic App Requires Best Performance

The initial algorithm was written in MATLAB, which is widely used in research and development by scientists and engineers who explore innovations. However, MATLAB is not optimized for commercial use when performance and efficiency are critical success factors.

To make the app commercially viable, Calpro required a faster, more mobile-friendly solution.

Our Computer Vision specialists rewrote the algorithm in C++ to improve performance. The new implementation was packaged as a library and integrated into both iOS and Android applications. This transition significantly enhanced execution speed, reduced computational overhead, and helped in seamless integration with mobile operating systems.

DAC.digital Team Composition

  • Computer Vision Architect
  • Computer Vision Engineer
  • Mobile Software Engineer
  • Delivery Manager
A graph representing migrating algorithm from Matlab to C++

Mockup App Helped Determine C++ Processing Speed

Before integrating the optimized algorithm into the final app, it was necessary to test its performance. We built a mockup application to compare execution times between the original MATLAB version and the new C++ implementation. The mockup app proved that C++ was the right choice for real-time computer vision in the app, as its processing speed improved what you can see in this table.

Now We’re Ready to Equip More Diagnostic Apps with Computer Vision

This project combined computer vision, mobile development, and high-performance C++ programming. While C++ is not our typical choice for mobile applications, we successfully implemented it to optimize the algorithm for speed and efficiency. The final app focused on smartphone-based test strip analysis, while our efforts were dedicated to developing the core computer vision algorithm and ensuring its seamless integration into the mobile ecosystem.

The Project Didn’t End Here: The App Is Getting New Features

We’re now starting the second phase of the project, where we are asked to expand the app’s functionality to include barcode scanning. The barcode contains metadata related to the test, so our algorithm now needs to process both the test strip and the barcode data.

Partner

Our partner was Apzumi who was responsible for developing the market-ready mobile app and invited our computer vision experts to this collaboration.

Got a Project in Mind? Let’s Discuss It.

Contact us!

Send us an email: [email protected]