Deep neural network (DNN) models can address these limitations of matrix factorization. DNNs can easily incorporate query features and item features (due to the flexibility of the input layer of the network), which can help capture the specific interests of a user and improve the relevance of recommendations.
DNN Face Detector in OpenCV
It is a Caffe model which is based on the Single Shot-Multibox Detector (SSD) and uses ResNet-10 architecture as its backbone. It was introduced post OpenCV 3.3 in its deep neural network module.
The image above shows our solution recognizing patients and people in religious events.
The solution prototype includes a smart watch that measures user vital health signs.
We designed a mobile app that is able to monitor vital health signs ( Blood pressure, Pulse rate, Body temperature, Oxygen level ) in addition to steps count for every user with the ability to retrieve data from 7 users at the same time using Bluetooth and send the collected data to the cloud for analysis.
Each smart bracelet is registered to the user ID. A smart glass will use AI to recognize the patient / user, retrieve his medical information stored on the cloud and show to the smart glass screen.
AI for Face recognition part
1) dnn.readNetFromCaffe is used to load the face detector
2) dnn.readNetFromTorch is used to the load face recognizer ( Reads a network model stored in TensorFlow framework's format. )
The code was part of a complete solution submitted in a healthcare competition and was nominated from the top 11 healthcare solutions. So using only a smart glass and smart watch, government personnel can monitor people's health in contactless settings.