Ovi Paul
Ph.D. Candidate at University of Houston
I am a Research Assistant at the National Center for Airborne Laser Mapping, University of Houston. I hold a bachelor’s degree in Computer Science and have previously worked as a Research Assistant and Machine Learning Engineer before beginning my Ph.D. studies. My research focuses on Computer Vision, Machine Learning, and Remote Sensing, particularly in developing intelligent models for geospatial understanding and environmental analysis.
BathyNet
Built a deep learning system in PyTorch using PointCNN to classify seafloor features from 3D LAS point-cloud data,
improving accuracy in bathymetric and underwater terrain mapping.
Text2Map
Developed an NLP and geospatial analysis pipeline that uses BERT and Hugging Face Transformers to extract and
geocode location entities from text, producing interactive spatial visualisations with GeoPandas and Folium
Adaptive-Retracker
A developing project that utilizes machine learning to analyze waveform altimetry data from various land types,
classify surface types, and perform adaptive retracking to enhance the accuracy of elevation estimates in complex terrains
Under-the-Sea
Created an interactive web platform in Streamlit to simulate sea-level rise impacts using elevation and census data
with PostGIS, GeoPandas, and Rasterio for large-scale spatial computations
During my academic journey I taught and assisted a variety of courses as a Teaching Assistant.
University of Houston
CIVE 3397 – Probability and Statistics for Engineers
Independent University, Bangladesh
CSC317 – Numerical Method
CSC204 – Data Structure
CSC101 – Introduction to Programming
Automated depth correction of bathymetric LiDAR point clouds using PointCNN semantic segmentation
Paul O, Ekhtari N, Glennie CL (2025). Frontiers Remote Sensing, 6:1521446.
BOLM high resolution land use and land cover dataset and benchmark results for the rapidly developing City of Dhaka Bangladesh
Paul O, Nayem ABS, Sarker A, et al. (2025). Nature Scientific Reports, 15:37689.
ROS-Pose Simplified object detection and planar pose estimation for rapid robotics application development
Paul SK, Paul O, Nicolescu M, Nicolescu M (2024). Software Impacts, 19:100624.
Image Pre-processing on NumtaDB for Bengali Handwritten Digit Recognition
Paul O (2018). International Conference on Bangla Speech and Language Processing (ICBSLP), Sylhet, Bangladesh, pp. 1–6.
