Human Activity Recognition Using Machine Learning Algorithms

CSE 445 - Machine Learning

In this project we used cellphone sensor data (angular motion, acceleration, geographic location and tri-axial (3-d) motion data) to classify human activity. The activity has been classified for 6 classes and a total of 3 machine learning models have been tested on this dataset. SVC performes with a highest testing accuracy of 95% followed by logistic regression with testing accuracy of 82% followed by Ridge classifier CV with testing accuracy of 81%. For further investigation about the models performance, we showed the precision, recall, f1-score along with corresponding confusion matrix.

SHOPLine | Augmented-Reality Based Cross Platform E-Commerce Application

Course : CSE327 - Software Engineering

The purpose of this project is to add a live camera 3D preview of the products that the customers buy from the online e-commerce sites.Our android version of the project will have this feature. User will just press a button and the camera will show a live 3d preview to the user. Which will make online shopping far more intresting and will hopefully attract more consumers.