Related Experiment Video
Updated: May 20, 2026
03:38
Unilateral Lung Volume Analysis Using Micro-CT for Enhanced Assessment of Pulmonary Fibrosis in Preclinical Models
Published on: June 20, 2025
Automatic annotation of histopathological images using a latent topic model based on non-negative matrix
Angel Cruz-Roa1, Gloria Díaz, Eduardo Romero
1BioIngenium Research Group, Faculty of Engineering and School of Medicine, Universidad Nacional de Colombia, Carrera 30 45-03 Ed 471 1er Piso, Bogotá D.C., 11001000, Colombia.
Journal of Pathology Informatics
|July 20, 2012
Summary
This study introduces a novel method for automatic histopathological image annotation using a bag of features, latent topic model, and probabilistic annotation. The approach significantly enhances the accuracy of identifying biological structures in medical images.
Area of Science:
- Digital Pathology
- Computational Biology
- Medical Image Analysis
Background:
- Histopathological images are crucial for clinical diagnosis and research.
- Automatic annotation of these complex images presents a significant challenge in image understanding.
Purpose of the Study:
- To develop and evaluate a novel, automated method for histopathological image annotation.
- To improve the accuracy and efficiency of extracting meaningful information from histopathological data.
Main Methods:
- A part-based image representation (bag of features) to capture fundamental biological patterns.
- A latent topic model using non-negative matrix factorization for high-level visual pattern discovery.
- A probabilistic annotation model linking visual features to 10 histopathological annotations.
Main Results:
- The method achieved a recall of 74% and a precision of 50% on 1,604 annotated skin tissue images.
- Demonstrated significant improvements over a baseline support vector machine method, with a 64% increase in recall and a 24% increase in precision.
Conclusions:
- The proposed method offers a robust and effective approach for automatic histopathological image annotation.
- This advancement has the potential to enhance diagnostic accuracy and accelerate biomedical research through improved image analysis.