Out-of-the-box code and models for CMU's object detection and tracking system for multi-camera surveillance videos. Speed optimized Faster-RCNN model. Tensorflow based. Also supports EfficientDet. WACVW'20

STEP: Spatio-Temporal Progressive Learning for Video Action Detection. CVPR'19 (Oral)

Code for our CVPR 2021 paper "Coarse-Fine Networks for Temporal Activity Detection in Videos"

End-to-End Semi-Supervised Learning for Video Action Detection [CVPR 2022]

[BMVC 2021]: Official PyTorch implementation of : "Few Shot Temporal Action Localization using Query Adaptive Transformers"

An example of how to adapt the C3D CNN for another problem. Really a fork of https://github.com/axon-research/c3d-keras

Acquire users location in android with ease, detect movement, geofence areas of interest, geocode users location

Internship in Computer Vision Lab at University of Louisville

This model detects workers' productivity states, using motion signals generated by the accelerometer embedded in the three mounted receiving beacons.

Activity Detection Using Unsupervised Learning Algorithms: DBSCAN, K-Means & Spectral Clustering to identify and label different activities in a dataset

ML application on wearable devices to predict a person's activities.

2-layer neural network that predicts exercise activity through IMU sensor data. For the graduate course "Introduction to Optimization and Machine Learning" at SJSU. Project partners, Antonio Cervantes and Christian Pedrigal.