I'm a robot vision researcher working on self-supervised learning for robot perception, aimed at reducing the need for labeled data in field and underwater environments. My past work includes benchmarking YOLOv11, RT-DETR, and Faster R-CNN for thermal-imaging-based crop detection, and I'm now pursuing a PhD in this space, with particular interest in agricultural and aquaculture robotics.
Benchmarked YOLOv11m against a QAAP-enhanced variant, Faster R-CNN, and RT-DETR for thermal capsicum detection, with a custom CLAHE-based preprocessing pipeline for low-quality thermal data.
Streamlit app for extracting structured data from invoice images using Qwen 2.5 VL via the OpenRouter API.
Automated PII detection and redaction for PDF documents using OCR and pattern recognition.
Attention-driven text-to-image synthesis using the AttnGAN architecture with word-level attention.
Real-time vehicle detection using YOLOv8, optimized for Raspberry Pi deployment.
CGPA: 3.17 / 4.00 (final two semesters: 3.65 / 4.00)