Motivation and values
- Machine learning models + IoT data = a smarter world (Google I/O ’18)
- To learn more about the IoT workflow
- To improve my skills in system engineering design and in programming
- To have a fun and meaningful project that can be a potential for bigger ideas
- To find the sweet spot where I can apply all of my skills, including machine learning
Project Description
- CycloPi is an IoT device that do image classification from a streaming video with cloud-based central monitoring and reporting
- CycloPi can act as an additional scanner at a self-service check out. For example, at Tesco whenever I buy banana or some certain type of grocery, I have to find them in the database. With CycloPi, it can immediately tell what I get and register it into the online inventory.
Why IoT Edge?
Edge computing means to process data close to where the things happen
- Quick response time and having access to the system even when the internet is down
- Use less network bandwidth
- Protect privacy
- Processing everything in the cloud can sometimes be too expensive
=> processes images from camera locally using machine learning and computer vision, then sends the processed data to the cloud to monitor, so the cameras act as a sensor providing the content of the images.
Results: