SharpAI Devkit

The world’s first deep learning embedded system development platform with AutoML.

Through the collection and upload of labled facial image samples, SharpAI Devkit uses AutoML to self-train and determine the best models for facial recognition. Throughout the collection and upload process, SharpAI recognizes faces within moving video, and continues to self-enhance it’s facial recognition database. This process enables high-reliability for multi-angle facial search and analysis.

Overview | Applications | Tech Specs

Product technology function

Innovative, easy to deploy AI solution

SharpAI uses AutoML to solve the typical challenges associated with training artifical intelligence systems. API interfaces enable fast and seamless customer integration.

Local architecture; enables low-cost and better security

At the core of its design, SharpAI considers the challenges associated with limited computational capabilities of embedded systems, and uses embedded ARM based devices to eliminate the high-cost of a dedicated GPU, providing a solid foundation for real-time video analysis.

Utilizes AutoML for optimized training

Using Point Circle ™, the training model automatically updates as new samples are collected. SharpAI becomes more effective over time as sample data increases.

Developer friendly

Through the REST API, SharpAI runs deep learning frameworks including TensorFlow and Caffe, which invoke related process through the interface.

Video Analysis Module

Through the search technology, DeepAI analyzes video streams and still images to enable real-time search for corresponding facial and object data. The app displays real-time search results which have been automatically filtered for relevancy, eliminating “false-positive” triggers.