Child Activity Recognition Using Machine Learning Techniques

Monitoring the behavior of children in the home is extremely important to avoid possible injuries. When it comes to monitoring children for providing better health facilities, Human Action Recognition (HAR) can be a valuable asset. Through HAR precise and resourceful monit

Project Title

Child Activity Recognition Using Machine Learning Techniques

Project Area of Specialization

Artificial Intelligence

Project Summary

Monitoring the behavior of children in the home is extremely important to avoid possible injuries. When it comes to monitoring children for providing better health facilities, Human Action Recognition (HAR) can be a valuable asset. Through HAR precise and resourceful monitoring of child action can be complex due to different views, temperament, resolution, and speed of motion of people, etc. Spatial-temporal information plays a vital role in human recognition. This project will focus on various models like CNN-RNN, Inflated I3D model etc. on the dataset and then compared to one another in order to predict accuracy and precision of results. Strategies of entire project will be written and executed in python following predictive analysis method to derive the results.

Project Objectives

Our main purpose it to detect the breast cancer at early stages

• Localization and tracking of the person within a video
• Feature extraction from the video
• Activity recognition, through the sequence of features in a time frame
• To monitor child activity using various machine learning models like CNN-RNN, Inflated I3D model.
• To evaluate the accuracy and efficiency of a combined machine and deep learning approach for child monitoring

Project Implementation Method

First step of our project will be Preprocessing. In preprocessing, we will do resizing the frame, trimming the video for the required frame and normalization. Second step will be Child detection and tracking. In this step, we will detect child from different objects. There could be other moving objects like any animal so we detection of child is necessary. To do this exact relevant portion from the video will be extracted. Third step will be Model selection. In this step we will be choosing which model we will be using in our project. CNN-RNN and I3D model will be used in our project. Next step we will be doing is Training our machine. After selecting our model, we will train our machine on different data sets. And the last step will be Evaluation. In evaluation, we will be calculating accuracy, training time, and loss value.

Benefits of the Project

• Localization and tracking of the person within a video
• Feature extraction from the video
• Activity recognition, through the sequence of features in a time frame
• To monitor child activity using various machine learning models like CNN-RNN, Inflated I3D model.
• To evaluate the accuracy and efficiency of a combined machine and deep learning approach for child monitoring

Technical Details of Final Deliverable

We will able to monitor our child even we are outside of the house. This will help us in our child's saftey.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Education

Other Industries

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Quality Education

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com