Coin Detection & Classification with Computer Vision

Gradient Insight partnered with an innovative fintech company to transform digital banking through a groundbreaking computer vision powered real-time coin recognition system. Picture 100% accuracy in coin detection, a 30% surge in operational efficiency, and an intuitive touchscreen interface redefining user satisfaction. Ready to revolutionize your fintech operations through the lens of computer vision? Explore the full case study for an in-depth look at the future of automated coin counting and precision in banking processes.


Gradient Insight collaborated with an innovative fintech company committed to redefining the future of digital banking. The project’s primary objective was to design and implement a real-time embedded system capable of recognizing and counting various types of coins, ultimately displaying the total amount to the user. This transformative solution aimed to enhance the efficiency and accuracy of coin counting processes in the banking industry.


The project presented several challenges including:

  • Accurate Coin Detection: Developing a precise object detection model to accurately identify and track different coin types in real-time under varying lighting conditions, angles, and coin orientations.

  • Real-Time Processing: Ensuring the system could recognize and count coins swiftly and accurately in real-time to provide an instant total amount to the user.

  • Embedded System Constraints: Designing a compact and efficient embedded system that meets the hardware limitations while maintaining high-performance standards.

  • Intuitive User Interface: Designing a user-friendly interface on the PiTFT touchscreen for real-time coin counts, amounts, and user commands (Start, Reset, Quit) required a delicate balance between functionality and simplicity.

Owl Flying Towards


Gradient Insight’s real-time coin recognition system revolutionized our fintech client’s operations, boosting efficiency by automating coin counting and allowing staff to prioritize complex tasks. The user-friendly touchscreen interface enhanced customer experience, while cost savings were realized through the elimination of manual counting, showcasing the system’s economic advantages.

The Solution

To tackle the challenges, Gradient Insight implemented a robust computer vision strategy. Accurate coin detection was achieved by deploying a customized object detection model, adept at identifying and tracking various coin types in real-time, even in challenging lighting conditions and diverse orientations. Real-time processing demands were met through optimization for swift and precise coin recognition, ensuring an instantaneous display of the total amount. Overcoming embedded system constraints involved efficient hardware design, balancing compactness with high performance. The creation of an intuitive PiTFT touchscreen interface was realized by carefully optimizing for functionality and simplicity, enabling users to seamlessly control the system and experience the full potential of real-time coin recognition in digital banking.

The Results


100% Recognition Accuracy

The object detection and tracking model ensured a reliable real-time coin recognition.

Enhanced User Satisfaction

The intuitive touchscreen interface received positive feedback, boosting overall user satisfaction.

30% Efficiency Boost

The embedded system reduced manual efforts and errors, leading to improvements in operational efficiency.

Sara Martins | CEO & Founder

Sara Martins | CEO & Founder

As the Director of Marketing and Operations at Gradient Insight, drives organizational success and growth by overseeing marketing efforts and optimizing operational activities.

Ready to harness the power of Data Science for your business?

Contact us today to learn more about how our team can help you achieve your data-driven goals!

Other Case Studies