Published October 31, 2025 | Version v1
Dataset Open

Gannan Grassland Rodent Burrow Detection Dataset (VOC format)

Authors/Creators

  • 1. ROR icon Lanzhou University

Description

This dataset contains rodent hole samples from the Gannan grassland region, formatted in the VOC standard. It includes a total of 42,400 images with corresponding labels. The training and validation sets are divided at a 9:1 ratio, and the combined training-validation set to test set ratio is also 9:1.

The dataset is accompanied by two pretrained model weight files:

  • CSPdarknet53_backbone_weights — the pretrained backbone network weights.

  • best_epoch_weights — the best-performing model weights obtained during training.

These weights are used in the CSPS_RetinaNet model, which can be downloaded from: https://github.com/yeshushahua/CSPS_RetinaNet.

Files

VOC_Mouse.zip

Files (2.0 GB)

Name Size
md5:dd49f63c0e1eba57654eb42585c1da06
146.8 MB Download
md5:2aacc07b79747c952170f187379269ab
106.7 MB Download
md5:a10717d205e5c8a06d5cb70eb72148b9
1.7 GB Preview Download