Evaluating Multiple Instance Learning Strategies for Automated Sebocyte Droplet Counting

2025-09-05

AI Summary

This study addresses the time-consuming and subjective nature of manual lipid droplet counting in sebaceous gland cell microscopy images by proposing an attention-based multiple instance learning (Attention-MIL) framework for automated, quantitative droplet enumeration. Methodologically, ResNet-50 extracts patch-level features, augmented with data augmentation, an MLP classifier, and a learnable attention pooling mechanism; performance is evaluated via five-fold cross-validation. Key contributions: (1) Simple bag-level aggregations (e.g., max/mean pooling) establish a robust baseline (mean MAE = 5.6); (2) Task-specific attention pooling design and regularization critically influence performance—some folds achieve superior accuracy (MAE = 3.2), yet overall stability remains limited (mean MAE = 10.7). To our knowledge, this is the first work applying MIL to sebum droplet quantification, revealing the pivotal role of pooling mechanisms and regularization in fine-grained biological counting tasks.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Multi-modal Vision
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Sebocytes are lipid-secreting cells whose differentiation is marked by the accumulation of intracellular lipid droplets, making their quantification a key readout in sebocyte biology. Manual counting is labor-intensive and subjective, motivating automated solutions. Here, we introduce a simple attention-based multiple instance learning (MIL) framework for sebocyte image analysis. Nile Red-stained sebocyte images were annotated into 14 classes according to droplet counts, expanded via data augmentation to about 50,000 cells. Two models were benchmarked: a baseline multi-layer perceptron (MLP) trained on aggregated patch-level counts, and an attention-based MIL model leveraging ResNet-50 features with instance weighting. Experiments using five-fold cross-validation showed that the baseline MLP achieved more stable performance (mean MAE = 5.6) compared with the attention-based MIL, which was less consistent (mean MAE = 10.7) but occasionally superior in specific folds. These findings indicate that simple bag-level aggregation provides a robust baseline for slide-level droplet counting, while attention-based MIL requires task-aligned pooling and regularization to fully realize its potential in sebocyte image analysis.

Problem

• Automating sebocyte lipid droplet counting to reduce manual labor
• Evaluating multiple instance learning for sebocyte image analysis
• Comparing baseline MLP with attention-based MIL model performance

Innovation

• Attention-based multiple instance learning framework
• ResNet-50 features with instance weighting
• Five-fold cross-validation benchmarking approach