Create your own conference schedule! Click here for full instructions

The Virtual Conference is located at

Abstract Detail

Conference Wide

Oberprieler, Christoph [1].

Deep Learning-Based Object Detection for Plant Scientists.

As more and more mature machine-learning methods for various types of automatization tasks become available, rich sources of hitherto underused biological image data are ready to be employed for quantitative analyses. In this workshop, participants will learn how to set up, train, and apply deep-learning models for object detection and segmentation for the localization and simultaneous classification of objects in digital images. The workshop will teach fundamental steps on the way from raw images to a trained model that can be used for making predictions on new data. As such, we will explore image annotation, image preprocessing, train-test splitting, and finally model training, evaluation and application. For this purpose we make use of the recent, enhanced version of the GinJinn [Ott et al. 2020, APPS] object detection framework. Practical applications will include localization and pixel-precise recognition (segmentation) of stomata in microscopic images, the identification and counting of herbivorous insects on sticky traps, the analysis of the composition of seed mixtures, and the extraction of leaf silhouettes from digitized herbarium specimens. After the workshop, participants will be able to apply deep-learning-based object detection and segmentation to their own data, without the need for manual programming.

Log in to add this item to your schedule

1 - Institute Of Plant Sciences, University Of Regensburg, Universitatsstr. 31, Regensburg, BY, D-93040, Germany

none specified

Presentation Type: Workshop
Session: W08, Deep Learning-Based Object Detection for Plant Scientists
Location: Virtual/Virtual
Date: Sunday, July 18th, 2021
Time: 10:00 AM(EDT)
Number: W08001
Abstract ID:19
Candidate for Awards:None

Copyright © 2000-2021, Botanical Society of America. All rights reserved