Computer Vision for Small-Object Perception
Tiny airborne object detection from moving platforms under clutter, camouflage, motion, and occlusion.
Md Hasibur Rahman
PhD Candidate in Computer Science, Missouri University of Science and Technology
Computer Vision · Machine Learning · Multimodal AI · Applied AI Engineering
I develop machine learning methods and applied AI systems for robust visual and multimodal perception under small-object scale, motion, occlusion, missing observations, sensor uncertainty, and domain shift. My dissertation work focuses on UAV-based airborne object detection and tracking, and I am extending these ideas toward GeoAI and biomedical AI.
Expected PhD: Fall 2026 · Open to postdoctoral research, research scientist, applied scientist, and AI/ML/CV engineering roles starting in 2026.
Research Interests
Spatiotemporal Tracking and Association
Temporal modeling, trajectory continuity, missing-observation recovery, and multi-object association.
Multimodal and Reliable AI
Learning from visual and multimodal signals under uncertainty, sensor degradation, missing inputs, and domain shift.
Multimodal UAV Sensing and Multi-LiDAR Perception
Robust sensing pipelines using heterogeneous LiDAR streams, temporal accumulation, clustering, tracking, and uncertainty-aware fusion for UAV-scale object perception.
Applied AI Engineering
Reproducible, testable, and deployment-aware AI pipelines for perception, evaluation, visualization, real-world data, and practical systems such as KeepScoreAI.
GeoAI and Biomedical AI
Remote sensing generalization and MRI/radiogenomic AI as extensions of robust multimodal learning.
Research Highlights
Representative systems from my UAV-based perception and tracking research.
STARD-Net
Spatiotemporal attention for detecting weak airborne targets from moving drones under clutter, motion, and occlusion.
KRAfT
Formation-aware multi-object tracking with Kalman residual refinement, conservative recovery, and missing-observation handling.
MultiLiDAR UAV Sensing
Heterogeneous LiDAR-based perception pipeline for sparse point clouds, temporal accumulation, candidate clustering, and reliability-aware fusion.
Selected Work
STARD-Net
Tiny airborne object detection from moving drones. ACM TSAS 2026.
KRAfT
Formation-aware UAV swarm tracking with conservative recovery. IEEE ICPR 2026 accepted.
MultiLiDAR UAV Sensing
Heterogeneous LiDAR-based UAV perception with temporal point-cloud accumulation and reliability-aware fusion. Ongoing research / applied AI engineering project.
KeepScoreAI
Applied AI engineering project for a practical AI-assisted scorekeeping workflow. Applied AI engineering project.
V-USDT
Vision-based UAV swarm detection and tracking using formation constraints. IEEE MDM 2025.
Augmented UAV Dataset
Dataset and augmentation work for UAV-based detection and tracking. IEEE AIPR 2023.
Radiogenomic MRI Classification
Biomedical AI extension on MRI-based radiogenomic classification. Pathways 2026 Top 3 Best Poster.
Recent Publications
- STARD-Net, ACM Transactions on Spatial Algorithms and Systems, 2026.
- KRAfT, IEEE ICPR, accepted, 2026.
- V-USDT, IEEE MDM, 2025.
Teaching & Mentoring
Teaching is an important part of my academic profile. I value clear explanation, reproducible implementation, and helping students connect computer science theory with working AI systems.
- Instruction and support across programming, data structures, algorithms, data management, AI, machine learning, and computer vision-oriented topics.
- Mentoring students on programming, debugging, applied ML workflows, and research implementation.
- Emphasis on first-principles explanation, reproducible experiments, and responsible use of AI tools.
Collaboration
I welcome collaboration and opportunities in computer vision, machine learning, multimodal AI, applied AI engineering, GeoAI, biomedical imaging, and reliable AI.