Blessing Agyei Kyem

I am a second year PhD student at the Sustainable Mobility and Advanced Research in Transportation (SMART) Lab at North Dakota State University, where I work on Multi-modal AI in Computer Vision, Agentic AI and their applications in Transportation and Pavement Engineering. My PhD advisor is Armstrong Aboah.

Email  /  GitHub  /  Google Scholar  /  ResearchGate  /  Blog  /  X  /  LinkedIn  / 

News

Research

I'm interested in deep learning, computer vision, LLMs, multimodal LLMs, and agentic ai.

Task-Specific Dual-Model Framework for Comprehensive Traffic Safety Video Description and Analysis

We propose a dual-model pipeline for traffic-scene video understanding that narrates safety-relevant events and answers questions about them using multimodal LLMs.

Prompt-Guided Spatial Understanding with RGB-D Transformers for Fine-Grained Object Relation Reasoning

We present a prompt-guided RGB-D multimodal LLM that performs fine-grained spatial reasoning in warehouse environments. It integrates object bounding boxes and depth cues into textual prompts to improve relation reasoning, distance estimation, and spatial grounding.

A Unified Detection Pipeline for Robust Object Detection in Fisheye-Based Traffic Surveillance

We propose a unified fisheye objection detection pipeline that combines fisheye-aware preprocessing, super-resolution/slicing postprocessing, and an ensemble of modern detectors to robustly detect small and distorted traffic objects.

Self-supervised multi-scale transformer with Attention-Guided Fusion for efficient crack detection

We present Crack-Segmenter, a self-supervised multi-scale transformer for pixel-level crack segmentation.

Saam-Reflectnet: Sign-Aware Attention-Based Multitasking Framework for Integrated Traffic Sign Detection and Retroreflectivity Estimation

This research involves designing a multi-tasking framework, called Saam-ReflectNet, that combines traffic sign detection, classification, and retroreflectivity estimation into a single workflow.

Context-CrackNet: A Context-Aware Framework for Precise Segmentation of Tiny Cracks in Pavement images

This study introduces Context-CrackNet, a novel architecture combining Region-Focused Enhancement Module (RFEM) and Context-Aware Global Module (CAGM) to improve crack segmentation.

Weather-Adaptive Synthetic Data Generation for Enhanced Power Line Inspection Using StarGAN

This research introduces a framework combining heuristic image processing and StarGAN to generate synthetic power line inspection images under adverse weather conditions like fog, rain, and nighttime.

PaveCap: The First Multimodal Framework for Comprehensive Pavement Condition Assessment with Dense Captioning and PCI Estimation

We show how to provide dense caption describing the condition of a pavement image. The textual description captures the types of pavement distress present with their severity, the absent pavement distresses and the pavement condition index (PCI)

Advancing Pavement Distress Detection in Developing Countries: A Novel Deep Learning Approach with Locally-Collected Datasets

This research presents a deep learning approach combining YOLOv5 with Convolutional Block Attention Module (CBAM) to detect and classify pavement distress types in developing countries.

Design and source code from Jon Barron's website