Monitoring Patients suffering from Alzheimers, Depression, Anxiety or any Mental Health Disease in Hospitals/Homes has been huge challenge as they can get self destructive. To solve this I am going to create Device using KV260 Vision AI Kit and OAK-D-IoT-40 to monitor patients remotely and track harmful behaviour and manage healthy daily routine.
Currently there are no automated solutions or any advanced monitoring solutions to track actual situation of patients unless use of human caretakers or doctors. My project uses Activity Tracking Models to monitor the patient and will alert the doctors and also predict such future aggressive behaviours or activities to the doctors. It will be much useful as 24x7 human monitoring wouldn't be require and it can be scaled easily with help of multiple cameras on single KV260 board.
The solution works by simply using 3D camera which is a Oak-D-40 Camera which can capture images with depth and it will send the frames to Xilinx Kria KV260 via USB for processing. The Activity Recognition model will be trained using Kinetics Dataset https://deepmind.com/research/open-source/kinetics .The custom model will be running on KV260 and whenever it recognises harmful activity from patient it will alert the doctors immediately. The 3D Camera is used because it can recognise more accurately what the patient is doing too avoid any false negatives.work by simply using 3D camera which is a Oak-F-40 Camera which can capture images with depth and it will send the frames to Xilinx Kria KV260 via USB for processing. The Activity Recognition model is trained using Kinetics Dataset https://deepmind.com/research/open-source/kinetics .The custom model will be running on KV260 and whenever it recognises harmful activity from patient it will alert the doctors immediately. The 3D Camera is used because it can recognise more accurately what the patient is doing too avoid any false negatives.
DepthAI library for Image Reading Kinetics Dataset
Comments
Please log in or sign up to comment.