Project Summary
This research proposes a Vision-to-Text system designed to automate the morphometric analysis of scuttle fly wings, addressing the limitations of manual identification which is currently labor-intensive and prone to error. By integrating Convolutional Neural Networks (CNNs) for precise landmark detection with Natural Language Generation (NLG), the study aims to extract quantitative data—such as wing length and costal indices—and transform it into standardized, text-based morphological descriptions. The project’s primary objectives are to develop this automated model, validate its accuracy against expert manual measurements, and ensure the generated reports meet taxonomic and forensic standards.
Researchers
Nurdanina Othman, Izfa Riza Hazmi & Raja M. Zuha