Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
April 3, 2026Scientific ReportsOpen Access

Coral morphology detection in underwater imagery using YOLOv12 with CNN and transformer encoder fusion

View Full Paper
Ask AI
Bookmark
Share

Authors

PNPriyanka NandalMSMeena SiwachGUGovind Murari Upadhyay

Discussion

Loading...

Member takes

Overview

Automated detection model enhances underwater image analysis for coral monitoring, indicating improved ecological insights.

Key Points

  • This work aims to develop an improved method for detecting coral morphology in underwater images using advanced machine learning techniques.
  • Utilized YOLOv12 for object detection.
  • Integrated Convolutional Neural Network for local feature extraction.
  • Applied transformer encoder for global context modeling.
  • Conducted empirical assessments on a standard underwater dataset.
  • Evaluated performance using precision, recall, and mean Average Precision.
  • Achieved higher accuracy compared to existing models like YOLOv7 and YOLOv11.
  • Demonstrated improvements in precision and recall metrics.
  • Maintained real-time inference capabilities.

Cite This Study

Nandal et al. (2026) studied this question.

synapsesocial.com/papers/69cf5c925a333a821460a284https://doi.org/10.1038/s41598-026-42591-z
View Full Paper
Ask AI
Bookmark
Share