Slide scanner-based microscopy image datasets of fossil diatoms from Southern Ocean surface sediments.

This dataset consists mainly of scanned images of permanent slides of surface sediments for the observation of fossil diatoms. The sediment samples were collected from a wide range of locations across the Southern Ocean. The permanent slides were prepared using standard processing, including hydrogen peroxide treatment and the pipette method. The image data include whole-coverslip scans (in the original format of Hamamatsu Photonics K.K.) and tile images (in JPEG format) segmented into 552 μm squares from the whole scanned area. All images have a resolution of 0.23 μm/pixel, which is sufficient for counting marine diatoms. The accompanying annotations (YOLO format) are specifically designed for detecting Eucampia antarctica (Castracane) Mangin, a key paleoenvironmental indicator. The tile images and annotations can be used to train and test object detection models for Eucampia antarctica, which can then be applied to morphometric analyses. The broad applicability of this dataset supports biogeographical and paleoenvironmental research through image-based analyses of fossil diatoms.