The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
MKProfiler (Maize-Kernel Profiler) v1.0
This Hugging Face repository provides a 500-photo benchmark subset for the MKProfiler maize-kernel phenotyping pipeline. The public processing release contains 15,278 valid ear images and approximately 0.38M kernels; this repository provides a smaller subset for inspection and reproducibility.
For the complete source code, environment setup, eight-stage workflow, model configuration, and usage instructions, please visit the main GitHub repository:
Contents
500p_test.rar— compressed archive containing 500 raw tray photographs.500p_test_results.rar— compressed archive containing the complete pipeline outputs for those photographs, including metadata, masks, contours, measurements, width profiles, latent traits, and final CSV tables.weights/— released YOLO, ResNet, VAE, and ONNX deployment assets.
Extract the two RAR archives to obtain the 500p_test/ and
500p_test_results/ directories.
The test images use the project imaging setup and are provided for pipeline inspection, downstream analysis, and reproducibility checks. The SAM2 checkpoint is not included; download it from the official SAM2 repository.
Recommended starting point
- Read the GitHub README.
- Download and extract
500p_test.rarand500p_test_results.rar. - Inspect
500p_test_results/metadata.csvand the final CSV files. - Use the released weights and the GitHub pipeline instructions if you want to reproduce the processing workflow.
License and contact
This pre-publication release is provided for research use under all-rights-
reserved terms. The licensing status is planned to be revisited after the
paper is published. Please contact Cedric Guo at 762323483@qq.com before
redistributing or using the data, code, or model files.
- Downloads last month
- 158