eyeM5 8500 edgeX HPC
- Realtime image saving in hard disc, and service provisioning of compression, decoding, and streaming dispatch.
- Highspeed real time image recognition and identification by adopting multiple types of AI model.
- AI Facial recognition with Openface architecture by capturing facial features and tracking the position of face.
- Vehicle plate recognition and identification provisioning based on the training of Caffe, the neural network framework, and Opencv DNN, a deep learning suite capturing the character of text to identify as a vehicle plate.
With new technological perspectives eyeball FinTech implemented AI Acceleration Engine onto the AI inside60 development platform. Differentiated from GUP’s accelerator, Xilinx FPGA chip architecture is adopted to solve the issue with respect to embedded AI distributed storage system by integrating AI acceleration engine and AI model zoo, which needs for highspeed computing, low energy consumption, e.g., 20w only, and efficient space occupation. This highly efficient development ecosystem with FPGA AI distributed storage technology has gotten good attention from around the world.
eyeM5 AI Image Recognition and Identification Embedded System, different from X86CPU + GPU system, is a subsystem of Xilinx Vitis, which is comprehensive with full range of R&D on FPGA-based chip, electric circuit, embedded Linux, FPGA-based AI IP models. eyeM5 AI Image Recognition and Identification Embedded System can empower clients for innovative solution development and experimentation, and has been deployed on highspeed rail in some Asian countries.
eyeM5 3500 AIOT Image Recognition and Identification Embedded System
The design and technological concepts are different from those of GPU-based platform design. The FPGA-based chip architecture can inherently tackle the issue of mounting embedded software modules to the framework of distributed storage platform that can fulfill the needs from the market for high computing power with good performance of low energy consumption (approximately 20w per hour), low latency, and low space required. eyeball FinTech developed an advanced AIOT image recognition and identification embedded system by integrating AI accelerating engine with AI IP model library.
eyeM5 AIOT Computer Lite
This microchip has big enough computing power to fulfill the requirement from applications to smart home, Internet of Vehicle, autonomous vehicle, etc. or data collection and analysis.