Theses and Dissertations

Date of Award


Document Type


Degree Name

Master of Science (MS)


Electrical Engineering

First Advisor

Dr. Weidong Kuang

Second Advisor

Dr. Heinrich Foltz

Third Advisor

Dr. Jingru Zhang


Detecting pedestrian flow in different directions at at traffic-intersection has always been a challenging task. Challenges include different weather conditions, different crowd densities, occlusions, lack of available data, and so on. The emergence of deep learning and computer vision algorithms has shown promises to deal with these problems. Most of the recent works only focus on either detecting combined pedestrian flow or counting the total number of pedestrians. In this work, we have tried to detect not only combined pedestrian flow but also pedestrian flow indifferent directions. Our contributions are, 1) we are introducing a synthetic pedestrian dataset that we have created using a videogame and a real-world dataset we have collected from the street. Our dataset has small, medium and high-density pedestrians crossing a crossroad, captured from different camera height, 2) We have proposed a Pedestrian Flow Inference Model (PFIM) that is trained on the synthetic dataset first and then is tested extensively on our real-world dataset. While testing on real-world dataset, we have embraced domain adaptation to reduce the domain gap between synthetic data and real-world data. Our proposed Pedestrian Flow Inference Model (PFIM) can detect pedestrian density and flow regardless of the height of the camera in three different ways - from left to right direction, from right to left direction, and total. Combining all, It has successfully tackled the challenges mentioned above and achieved state-of the-art performances.


Copyright 2021 Abhijit Baul. All Rights Reserved.