Suraj Maniyar

Deep-Visual-Attention-Prediction. Keras implementation of paper 'Deep Visual Attention Prediction' which predicts human eye fixation on view-free scenes.

Stock-Trading-Using-Machine-Learning. A comprehensive approach for stock trading implemented using Neural Network and Reinforcement Learning separately.

Time-Series-Prediction. Exploring HMM, LSTM and Regression techniques to predict respiratory rate of an individual from accelerometer data.

Activity-Recognition-From-Video. A deep learning model utilizing CNN and LSTM to recognize activity from video. (To be used for bench-marking hardware accelerator)

Single-View-Metrology. 3D construction of an object from a 2D image using 3 point perspective.

Monocular-SLAM-AutonomousNavigation-ECE592. Implementation of VINS-Mono SLAM algorithm on NVIDIA Jetson TX1 to map a construction site and localize an aerial robot blimp.

Foraminifera-Image-Segmentation. Biological Image Segmentation from edge probability map using Graph-Cut and Watershed algorithm

Siamese-CNN-MNIST. Digit classification using Siamese CNN.

Task-Learning-Robot. An attempt to implement 'Robot Learning from Demonstration' where we use industrial robotic arm Scorbot-ER VII to perform a desired task.

Object-Detection-SSD300. Object tracking using SSD

VQA-PyTorch. My implementation of VQA in PyTorch