This deep learning library for embedded devices was developed by Uncanny Vision (Head Office: India).
This product is optimized for Arm NEON and makes it possible to deploy advanced deep learning technology even in embedded devices with strict resource constraints, such as ADAS (Advanced Driver Assistance Systems), machine vision, robots, surveillance cameras, and mobile apps.
• Uncanny Vision has a long and successful track record in developing algorithms for embedded devices. Uncanny Vision developed this product by using Convolutional Neural Network (CNN), a method for implementing deep learning
• Optimized for Arm NEON (for Arm Cortex-A series processors based on Arm NEON)
• Deep learning, which requires very large amounts of computational processing, can be used in embedded devices without using the GPU or cloud computing
• ADAS (Advanced Driver Assistance Systems)
• Machine vision
• Surveillance cameras
• Mobile apps
What is deep learning?
• This technology enables advanced recognition of data such as audio, images, and text by using a neural network, which is a learning model that imitates the function of the human brain.
• CNN is one method of deep-learning implementation, which improves on previous neural networks designs.
• CNN has a multilayered structure consisting of an input layer, multiple intermediate layers (hidden layers), and an output layer. CNN extracts characteristics in stages for lines and points, parts, and overall images, and combines these elements to display the results.
• When groups of target images are learned in advance, the displayed results become more accurate.
Deep-learning library evaluation kit
Super high-speed computer vision library
for ARM NEON based processors
PUX image recognition software