Quantitative assessment of osteoclasts by counting and measuring area proportions is a fundamental approach for analyzing in vitro osteoclastogenesis. However, there is a lack of user-friendly automated analysis tools that do not require programming expertise, leading researchers to rely on manual operations for quantitative analysis. To address this limitation, the present study developed an offline software application utilizing the YOLOv8 instance segmentation model. This software facilitates the automated detection of osteoclasts in single images and batches, achieving an inference speed of less than one second per image. Following detection, the system promptly generates a visualized prediction report for user reference, providing essential data support for subsequent analyses. To ensure optimal model performance, the model was trained on a high-quality dataset meticulously annotated by experts in bone histology. The model ultimately achieved a mean average precision of 95.3% in detecting osteoclasts derived from mouse bone marrow macrophages, facilitating swift feedback of quantitative analysis results and significantly enhancing analysis efficiency. In conclusion, this study developed an accessible quantitative analysis tool for osteoclasts, intended for researchers without programming expertise. • An automated quantitative analysis system for osteoclasts has been developed, enabling precise quantification of osteoclast number and area. This system demonstrates high accuracy and rapid detection, significantly improving research efficiency. • The model is packaged as a standalone offline application, enabling researchers without programming experience to use it effortlessly. • The YOLOv8l-seg model was evaluated on both segmentation and detection tasks to verify its performance.
Huang et al. (2026) studied this question.