As a demonstration of an AI integration solution for embedded devices, we built a Deep Learning application that runs on Raspberry Pi 3.
We have built an application system using Deep Learning that operates on a Raspberry Pi 3 as a proposal for AI solutions for edge devices. At Embedded Technology West 2017 / the Kansai Embedded Technology Exhibition, we conducted a technical demo of a celebrity look-alike judgment app using PiCamera. Anticipating various future needs and versatility, we created it from scratch on AWS using a Deep Learning framework without utilizing existing pre-trained models. We provide technologies and ideas that can flexibly adapt to and control the embedded devices envisioned by our customers. *For more details, please refer to the PDF document or feel free to contact us.
Inquire About This Product
basic information
【System Contents】 ■dlib ・Face recognition library ・Used for extraction from material data (face photos) ■chainer ・Deep Learning framework ・Adopted chainer for its ease of use and simplicity ■AWS ・Training models are trained on AWS GPU instances ・Using p2.xlarge on Amazon EC2 P2 instances ■PyQt4 ・Used for GUI creation ■PiCamera ・Camera library for Pi ・Used for camera shooting on the device ■Anaconda3-4.0/Miniconda3.4 ・Pytorch distribution ・For Raspberry Pi 3, configured minimally with Miniconda ■RaspberryPi3/Raspbian ・Keeps the training model locally and enables all operations locally *For more details, please refer to the PDF document or feel free to contact us.
Price range
Delivery Time
Applications/Examples of results
For more details, please refer to the PDF document or feel free to contact us.
catalog（1）
Download All Catalogs
Company information
At Bee, we handle not only embedded software development but also the development of associated applications and cloud/AI solutions in a comprehensive manner, following innovative technologies that are not confined to traditional embedded software frameworks. We are certified under the information security management system standard "ISO/IEC 27001:2013" in customer requirements as well as in the design, development, and maintenance of our proprietary software. Furthermore, while "DX transformation" is being emphasized in recent times, we have achieved paperless and cloud-based operations since our founding in 2006, considering improvements in both information security and the internal environment as standard practice.
