Optical topometry and machine learning to rapidly phenotype stomatal patterning traits for maize QTL mapping

Dataset

Description

This dataset containes the images of B73xMS71 RIL population used in QTL linkage mapping for maize epidermal traits in year 2016 and 2017.
2016RIL_all_mns.rar and 2017RIL_all_mns.rar: contain raw images produced by Nanofocus lsurf Explorer Optical Topometer (Oberhausen, Germany) at 20X magnification with 0.6 numerical aperture. Files were processed in Nanofocus μsurf analysis extended software (Oberhausen,Germany).
2016RIL_all_TIF.rar and 2017RIL_all_TIF.rar: contain images processed from the Topology layer in each nms file to strengthen the edges of cell outlines, and used in downstream cell detection.
2016RIL_all_detection_result.rar and 2017RIL_all_detection_result.rar: contain images with epidermal cells predicted using the Mask R-CNN model.
training data.rar: contain images used for Mask R-CNN model training and validation.
Date made availableJul 10 2021
PublisherUniversity of Illinois Urbana-Champaign

Keywords

  • water use efficiency
  • cell segmentation
  • stomata
  • Mask R-CNN

Cite this