@article{ART003306417},
author={Tae Woo Kim and Gyeongmin Kim},
title={A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition},
journal={Journal of Internet of Things and Convergence},
issn={2466-0078},
year={2026},
volume={12},
number={1},
pages={139-145}
TY - JOUR
AU - Tae Woo Kim
AU - Gyeongmin Kim
TI - A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition
JO - Journal of Internet of Things and Convergence
PY - 2026
VL - 12
IS - 1
PB - The Korea Internet of Things Society
SP - 139
EP - 145
SN - 2466-0078
AB - Marine ecosystem surveys record benthic community information such as percent cover within a standardized area using photoquadrat based surveys, but automation is required because interior delineation and annotation are repeated. This study proposes a two stage pipeline that combines object detection with prompt based segmentation to delineate the quadrat interior in underwater quadrat images. In Stage 1, a YOLO11 based detector estimates the quadrat bounding box. In Stage 2, the estimated box is used as a box prompt for AquaSAM to generate an interior mask. AquaSAM was fine tuned on the SUIM dataset. When the integrated pipeline was applied, masks were generated for most test images, but continuity and completeness decreased under seaweed occlusion, reflections, and low contrast, and partial omissions occurred. This demonstrates feasibility, but improved robustness and broader quantitative evaluation are needed to reduce quality variation across scene conditions.
KW - Underwater images;Quadrat recognition;Object detection;Prompt-based segmentation; AquaSAM;YOLO11
DO -
UR -
ER -
Tae Woo Kim and Gyeongmin Kim. (2026). A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition. Journal of Internet of Things and Convergence, 12(1), 139-145.
Tae Woo Kim and Gyeongmin Kim. 2026, "A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition", Journal of Internet of Things and Convergence, vol.12, no.1 pp.139-145.
Tae Woo Kim, Gyeongmin Kim "A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition" Journal of Internet of Things and Convergence 12.1 pp.139-145 (2026) : 139.
Tae Woo Kim, Gyeongmin Kim. A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition. 2026; 12(1), 139-145.
Tae Woo Kim and Gyeongmin Kim. "A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition" Journal of Internet of Things and Convergence 12, no.1 (2026) : 139-145.
Tae Woo Kim; Gyeongmin Kim. A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition. Journal of Internet of Things and Convergence, 12(1), 139-145.
Tae Woo Kim; Gyeongmin Kim. A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition. Journal of Internet of Things and Convergence. 2026; 12(1) 139-145.
Tae Woo Kim, Gyeongmin Kim. A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition. 2026; 12(1), 139-145.
Tae Woo Kim and Gyeongmin Kim. "A Two-Stage Deep Learning Pipeline for Underwater Quadrat Recognition" Journal of Internet of Things and Convergence 12, no.1 (2026) : 139-145.
