Neural-Aggregation-Network-for-Video-Face-Recognition

This repo is an implementation of 'Neural Aggregation Network for Video Face Recognition' by Jiaolong Yang. The basic idea behind it is to represent multiple features of the same subject as a single aggregated feature. While aggregation images which are clearer gets a higher weight as compared to ones that are blur. A model is trained for this purpose.

Human-Activity-Recognition-using-CNN

CNN model that classifies human activity with a test accuracy of 91.6%. The target labels are 'walking,' 'jogging', 'going upstairs', 'going downstairs', 'sitting' and 'standing'. Data was collected from UCI Machine Learning Repository and each activity is recorded in m/s^2 along 3-axis.