Case Study - Incloud background
Case Study - Incloud background
Case Study - Incloud logo

Keeping the vision to make development a painless experience for all clients in mind, InCloud always wants to deliver web or mobile applications, IOT SENSOR SYSTEMS, and CLOUD SERVICES CONSULTING with Japanese-level quality with Western-level technological skills at very cost-effective prices.

Case Study - Incloud background
Case Study - challenges icon

Challenges

InCloud required a development team with qualified AI engineers not only with high expertise and experience in multiple AI technologies but also flexible to scale up and down, depending on the project’s situation and workload.

Recommended by Orient, a loyal partner of InCloud since the very first days via many projects such as Smart QA or Ethera, InCloud decided to give trust in Neurond AI services to help them realize their idea of an AI-powered seven segment digital display detection system.

Case Study - Incloud background
Case Study - Solutions icon

Solution

To meet InCloud's demand, Neurond brought together a dedicated team of four AI engineers. With solid experience in AI and the ability to work under pressure, they’ve greatly contributed to the success of the Athena project.

Our team also worked closely with InCloud to discuss, update on the progress, and receive feedback on every small stage of the project. As AI projects need well preparation at the planning stage, trying out different models for the most effective one, we had to invest a lot in researching and analyzing. Our AI engineer looked carefully at the accuracy and security to make sure the system performed correctly.

This analog device digitalization system applied Computer Vision technology, using Custom Single Shot Multi-box Detection, Screen Localization using YOLO, Transfer to Lite Model & Quantized, Model to Run on Raspberry Pi v4 (30fps), and Confirmation Method. The AI model ran on Tensorflow Lite and could be run entirely offline on a Raspberry Pi or any other Single Board Computer. The system can be connected via HTTP to our inbuilt GUI or via UDP/HTTP API to your own application.

Technologies Used

  • Computer Vision
    • Custom Single Shot Multi-box Detection
    • Screen Localization
Case Study - Incloud background
Case Study - Impact icon

Impact

Neurond's dedicated AI team greatly contributed to the success of the Athena project. It offers extremely accurate, lightning-fast detection, and a completely offline solution for detecting 7-segment digital displays.

The platform reaches a 99.8% accuracy and 5ms detection rate under ideal conditions. It can even be trained for various angles and lighting conditions. Plus, customers can use it in both online and offline modes.

Following the success of Athena, InCloud continued working with Orient on other AI projects like X-ray AI.