PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 6, 2026Modelling—International Open Access Journal of Modelling in Engineering Science0 citationsOpen Access

A SAM2-Driven RGB-T Annotation Pipeline with Thermal-Guided Refinement for Semantic Segmentation in Search-and-Rescue Scenes

View Full Paper
ASAndrés Salas-EspinalesRVRicardo Vázquez-MartínAMAnthony Mandow

Key Points

  • The central aim is to develop a reliable RGB-T annotation pipeline for semantic segmentation in search-and-rescue applications.
  • Developed a semi-automatic annotation pipeline using SAM2 in Label Studio.
  • Integrated deep correspondence matching and geometric transformation for RGB-T pairs.
  • Implemented guided human refinement and quantitative quality control through inter-annotator agreement.
  • Reduced annotation time by 36% with high annotation quality (mean IoU = 74.9%).
  • Achieved strong inter-annotator agreement (mean pixel accuracy = 74.3%, Cohen’s κ = 65%).
  • Annotated a dataset of 306 RGB-T image pairs, which was used for training state-of-the-art models.

Abstract

High-quality RGB–thermal infrared (RGB-T) semantic segmentation datasets are crucial for search-and-rescue (SAR) applications, yet their development is hindered by the scarcity of annotated ground-truth and by the challenges of thermal-camera calibration, which typically depends on heated targets with limited geometric definition. Recent approaches focus on using semantic segmentation annotation tools and transferring RGB masks to multi-spectral data, but they do not fully address the need for robust cross-modal geometric validation, quality control, or human-in-the-loop reliability assessment in RGB-T segmentation. To fill this gap, we propose a validated cross-modal annotation pipeline that combines deep correspondence matching, geometric transformation (affine or homography) of RGB-T pairs, and quantitative alignment validation. Our RGB-T pipeline integrates a semi-automatic annotation pipeline based on the Segment Anything Model 2 (SAM2) in Label Studio, with guided human refinement, and incorporates quantitative cost and quality control via inter-annotator agreement before being used in downstream model training. Results across three annotators show that the proposed approach reduces annotation time by 36% while achieving high annotation quality (mean IoU = 74.9%) and strong inter-annotator agreement (mean pixel accuracy = 74.3%, Cohen’s κ = 65%). The proposed RGB-T pipeline was annotated on a SAR-oriented RGB-T dataset comprising 306 image pairs and trained on two SOTA RGB-T. These findings demonstrate the practical value of the proposed methodology and establish a reproducible framework for generating reliable RGB-T semantic segmentation datasets, complementing and extending recent multispectral auto-labeling approaches.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Salas-Espinales et al. (2026) studied this question.

synapsesocial.com/papers/69aa7096531e4c4a9ff5a866https://doi.org/10.3390/modelling7020050
Ask AI
Helpful
Bookmark
Share
View Full Paper