Few-shot rock type classification without slide labels using polarized light pairing
Cross-polarization pairing for label-free few-shot thin-section classification
Computer Vision and Pattern Recognition
Summary
Identifying rock types in microscope images usually needs experts to label each slide, which is slow and expensive. The authors developed a way to teach a computer to recognize new rock types without needing those labels, by using pairs of images taken under two different light conditions that show the same place on the slide. This pairing helps the computer learn shapes and textures instead of just colors, improving accuracy. Their method performs almost as well as models trained with many labeled examples.
What this means in practice
- •For petrographers: Classify new rock thin section images without costly manual labeling by using paired polarized light views to improve recognition accuracy.
- •For geological survey teams: Automate identification of diverse rock types from microscope images in field or lab settings when labeled training data is unavailable.
Authors
Bingchen Yan
Abstract
Few-shot classifiers of rock thin sections are meta-trained on many rock types that a petrographer has identified slide by slide, and this identification is the costly step that new collections lack. We study few-shot recognition of unseen rock types when no training slide is identified, and the only information available is what the microscope records anyway: which slide an image comes from, and whether it was taken in plane-polarized (PPL) or cross-polarized (XPL) light. We first show why generic unsupervised pseudo-tasks fit this setting poorly: the positives they find without labels are either wrong or near-duplicates of the anchor, so they teach the model little. Polarized-light microscopy provides positives that avoid both problems. The PPL and XPL views of one field show the same grains, so they are correct by construction, yet they look very different. We propose cross-polarization pairing (XPP), which treats each slide as a pseudo-class and pairs its PPL and XPL views of the same field location, so that matching must rely on grain shape and texture rather than on colour. On the Nanjing University collection of 108 rock types, with support and query images taken from different slides, XPP is the best label-free method on average over three random splits, and matches a supervised model trained on 78 identified rock types. Controls show that colour augmentation and other acquisition positives do not reproduce the gain. We also show that the image-level protocol of earlier work is saturated, and propose a cross-slide protocol in its place.