Evaluating Multiple Instance Learning Strategies for Automated Sebocyte Droplet Counting
This work addresses the labor-intensive and subjective manual counting problem for researchers in sebocyte biology, but it is incremental as it benchmarks existing methods without introducing major innovations.
The study tackled automated counting of lipid droplets in sebocyte cells by comparing a baseline multi-layer perceptron with an attention-based multiple instance learning model, finding that the baseline achieved more stable performance with a mean MAE of 5.6 versus 10.7 for the MIL model.
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.