Related Experiment Video
Updated: Jun 12, 2026
05:33
Using Robotic Systems to Process and Embed Colonic Murine Samples for Histological Analyses
Published on: January 7, 2019
6.0K
Improving the Annotation Process in Computational Pathology: A Pilot Study with Manual and Semi-automated Approaches
Giorgio Cazzaniga1, Fabio Del Carro1, Albino Eccher2
1Department of Medicine and Surgery, Pathology, IRCCS Fondazione San Gerardo Dei Tintori, University of Milano-Bicocca, Via Pergolesi, 33, 20900, Monza, Italy.
Journal of Imaging Informatics in Medicine
|September 4, 2024
Summary
Semi-automated annotation using Segment Anything Model (SAM) significantly speeds up pathology image analysis and improves reproducibility compared to manual methods. This AI-assisted approach enhances ground truth for developing reliable artificial intelligence algorithms in digital pathology.
Area of Science:
- Digital Pathology
- Artificial Intelligence in Medicine
- Computational Pathology
Background:
- Developing reliable artificial intelligence (AI) algorithms in pathology requires accurate ground truth data from whole slide image (WSI) annotations.
- Manual annotation of WSI is a time-consuming and operator-dependent process, hindering efficient AI development.
- Exploring streamlined annotation approaches is crucial for advancing AI in pathology.
Purpose of the Study:
- To compare the efficiency, reproducibility, and precision of different annotation methods for renal tissue in whole slide images.
- To evaluate the performance of a semi-automated approach (Segment Anything Model - SAM) against manual methods (mouse and touchpad).
- To assess the impact of annotation tools and operator variability on ground truth generation for AI development.
Main Methods:
- Two pathologists annotated renal tissue compartments (tubules, glomeruli, arteries) using semi-automated (SAM) and manual (mouse, touchpad) tools.
- Key metrics included working time, reproducibility (overlap fraction), and precision (accuracy rating).
- The influence of different display monitors on mouse annotation performance was also analyzed.
Main Results:
- The semi-automated SAM approach was significantly faster (13.6 min) and showed less inter-observer variability (2% difference) than mouse (29.9 min, 24% difference) and touchpad (47.5 min, 45% difference).
- SAM achieved the highest reproducibility for tubules (1.0) and glomeruli (0.99), while mouse and touchpad showed slightly lower values.
- No significant precision differences were found between operators (p=0.59), but non-medical monitors increased annotation time by 6.1%.
Conclusions:
- Semi-automated and AI-assisted annotation methods, like SAM, can substantially accelerate the ground truth generation process in digital pathology.
- These advanced approaches offer improved efficiency and reproducibility, crucial for the development of robust AI tools.
- Optimizing annotation workflows is key to enhancing the reliability and scalability of AI in pathological diagnostics.
Keywords:
AnnotationArtificial intelligenceComputational pathologyDigital pathologySegment Anything ModelMore Related Videos
Related Concept Videos
02:07Blind Procedures
Ideally, the people who observe and record the children’s behavior are unaware of who was assigned to the experimental or control group, in order to control for experimenter bias. Experimenter bias refers to the possibility that a researcher’s expectations might skew the results of the study. Remember, conducting an experiment requires a lot of planning, and the people involved in the research project have a vested interest in supporting their hypotheses. If the observers knew which child was...
01:30Equipments Used To Measure Blood Pressure
Direct Method
This invasive approach involves cannulating a peripheral artery. During each cardiac contraction, pressure generates mechanical motion within the catheter, transmitted through rigid, fluid-filled tubing to a transducer. This transducer converts mechanical motion into electrical signals displayed as waveforms on a monitor. An automatic flushing system prevents blood backflow. Due to the potential risk of unexpected arterial blood loss, this method is primarily used in intensive...
This invasive approach involves cannulating a peripheral artery. During each cardiac contraction, pressure generates mechanical motion within the catheter, transmitted through rigid, fluid-filled tubing to a transducer. This transducer converts mechanical motion into electrical signals displayed as waveforms on a monitor. An automatic flushing system prevents blood backflow. Due to the potential risk of unexpected arterial blood loss, this method is primarily used in intensive...
01:23Types of Biopharmaceutical Studies: Controlled and Non-Controlled Approaches
Biopharmaceutical studies constitute a vital field aiming to enhance drug delivery methods and refine therapeutic approaches, drawing upon diverse interdisciplinary knowledge. In research methodologies, the choice between controlled and non-controlled studies significantly influences the study's reliability and accuracy.
Non-controlled studies, commonly employed for initial exploration, lack a control group, rendering them susceptible to biases and external influences. In contrast, controlled...
Non-controlled studies, commonly employed for initial exploration, lack a control group, rendering them susceptible to biases and external influences. In contrast, controlled...