Learn how to build multiple computer vision inference neural networks that unobtrusively gather data on student behavior and activity in the classroom.
Gain insight into the following solutions:
Computer vision applications for IoT
Inference to analyze video data
Visual solutions market in IoT
Learn to build and run an application with these capabilities:
❶ Detect the attentiveness of students in the classroom through head-pose estimation inference models
❷ Capture real-time classroom activity and determine attendance through action detection and facial re-identification
❸ Visualize the output in a data-rich dashboard
How It Works
The intelligent classroom solution demonstrates how to combine several neural networks to detect real-time classroom metrics.
The application marks students looking toward the front of the class as attentive. To calculate this metric, the application uses head-pose-estimation-adas-0001 to get the value of the head pose of each student and calculate the mean.
To capture real-time class participation metrics (such as students standing or raising a hand), it uses person-detection-action-recognition-0005.
To detect the average emotions of the students and calculate the overall mood of the classroom in real time, it uses the pretrained model emotions-recognition-retail-0003.
To calculate the total class attendance, it uses the pretrained model face-reidentification-retail-0095 to detect individual faces.