Youtor
From CES 2020 Artificial Intelligence Developers Lab
>> YOUR LINK HERE: ___ http://youtube.com/watch?v=Wzgfm5BH4qY
Find out more information: http://bit.ly/CES20-ai • Face Identification • Detection and identification of multiple faces with bounding box displayed in real-time from a live video feed. • • Enrollment via PC of up to 8 new faces • • Preprocessing, face detection, image cropping and scaling are performed by the STM32H747 • Face Expression Recognition • Live face expression recognition from a camera video stream. • • Detection of multiple faces with labeled bounding box displayed in real-time • • Recognizes 7 expressions: angry, disgusted, scared, happy, neutral, sad, surprised • • Preprocessing, face detection, image cropping and scaling are performed by the STM32H747 • Image Classification on Microprocessor • Advanced classification of images among 1000 different categories using Tensor Flow Lite on the STM32MP1 dual-core MPU. • • Images are classified with MobileNet v1: 1000 classes at 8 frames per second • • Tensor Flow Lite integrated via C++ runtime implementation on dual-core A7 • • Avenger96 board with parallel interface Camera daughter board. Allows optimal image capture without USB overhead • Contextual Activity Recognition • Distributed intelligence between ST Smart Sensor Machine Learning Core and a neural network running on STM32 microcontroller. • • Ultra-low-power always-on Human Activity Recognition on LSM6DSOX Machine Learning Core • • Low-power sound analysis neural network running on STM32L4 triggered by the activity recognition • • Audio and motion are combined to recognize 5 activities (stationary, walking, running, biking, driving) in 3 different contexts (outdoor, indoor or in vehicle) • Condition Monitoring of a Multi-Fan System • NanoEdge AI Studio by Cartesiam easily creates custom machine learning library able to learn and infer in STM32 microcontrollers. No data set, no data scientist, no pre-trained Artificial Neural Network required. • • Vibration analysis in a multi-source and noisy environment • • On-chip learning of the normal condition • • On-chip anomaly detection such as shock, clogging, etc. is displayed real-time
#############################
