Fine-Tuning Vision Classifiers on a Budget

Abstract

Fine-tuning modern computer vision models requires accurately labeled data for which the ground truth may not exist, but a set of multiple labels can be obtained from labelers of variable accuracy. We tie label quality to confidence derived from historical labeler accuracy using a simple naive-Bayes model. Imputing true labels in this way allows us to label more data on a fixed budget without compromising label or fine-tuning quality. We present experiments on a dataset of industrial images that demonstrates that our method, called Ground Truth Extension (GTX), enables fine-tuning ML models using fewer human labels.