Deep Learning for Multi-task Plant Phenotyping

@article{Pound2017DeepLF,
  title={Deep Learning for Multi-task Plant Phenotyping},
  author={Michael P. Pound and Jonathan A. Atkinson and Darren M. Wells and Tony P. Pridmore and Andrew P. French},
  journal={2017 IEEE International Conference on Computer Vision Workshops (ICCVW)},
  year={2017},
  pages={2055-2063}
}
Plant phenotyping has continued to pose a challenge to computer vision for many years. There is a particular demand to accurately quantify images of crops, and the natural variability and structure of these plants presents unique difficulties. Recently, machine learning approaches have shown impressive results in many areas of computer vision, but these rely on large datasets that are at present not available for crops. We present a new dataset, called ACID, that provides hundreds of accurately… 

Figures and Tables from this paper

Doing More With Less: A Multitask Deep Learning Approach in Plant Phenotyping
TLDR
This work presents a multitask deep learning framework for plant phenotyping, able to infer three traits simultaneously: leaf count, projected leaf area, and genotype classification, using a modified pretrained ResNet50 as a feature extractor.
Convolutional Neural Networks for Image-Based High-Throughput Plant Phenotyping: A Review
TLDR
The goal of this review is to provide a comprehensive overview of the latest studies using deep convolutional neural networks (CNNs) in plant phenotyping applications, and specifically review the use of various CNN architecture for plant stress evaluation, plant development, and postharvest quality assessment.
SpikeSegNet-a deep learning approach utilizing encoder-decoder network with hourglass for spike segmentation and counting in wheat plant from visual imaging
TLDR
A dedicated deep learning approach has been developed to identify and count spikes in the wheat plants based on combined digital image analysis and deep learning techniques, and the performance of the approach demonstrates that SpikeSegNet is an effective and robust approach for spike detection and counting.
Towards Automated Analysis of Grain Spikes in Greenhouse Images Using Neural Network Approaches: A Comparative Investigation of Six Methods
TLDR
A comparative investigation of six neural network methods for pattern detection and segmentation in RGB images, including five deep and one shallow neural network shows superior performance, and a GUI-based tool is provided, which shows the application of pre-trained neural networks to fully automate spike detection, segmentation and phenotyping in images of greenhouse-grown plants.
Towards Low-Cost Image-based Plant Phenotyping using Reduced-Parameter CNN
TLDR
This work re-architects four baseline deep neural networks by reducing their parameters whilst making them deeper to avoid the problem of overfitting, and introduces a simple global hyper-parameter (alpha) that provides an efficient trade-off between parameter-size and accuracy.
Computer vision-based phenotyping for improvement of plant productivity: a machine learning perspective
TLDR
Computer vision-based phenotyping will play significant roles in both the nowcasting and forecasting of plant traits through modeling of genotype/phenotype relationships.
Application of deep learning and machine learning workflows for field-scale phenotyping
TLDR
A high throughput phenotyping approach that is based on extracting time-lapse information of anthesis progress from digital cameras is utilized to create a robust algorithm for tassel phenotypes.
Visualizing Feature Maps for Model Selection in Convolutional Neural Networks
  • Sakib Mostafa
  • Computer Science
    2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW)
  • 2021
TLDR
Guided Backpropagation is used to show how the pairwise similarity matrix between the visualization of the features learned at different depths can be leveraged to pave a new way to potentially select a better CNN model by removing redundant layers.
Web-SpikeSegNet: Deep Learning Framework for Recognition and Counting of Spikes From Visual Images of Wheat Plants
TLDR
An online platform based on a deep-learning framework for spike detection and counting from the wheat plant’s visual images, Web-SpikeSegNet is a significant step forward in the field of wheat phenotyping and will be very useful to the researchers and students working in the domain.
WTPlant (What's That Plant?): A Deep Learning System for Identifying Plants in Natural Images
TLDR
First experiments support the hypothesis that an initial segmentation process helps guide the extraction of representative samples and, consequently, enables Convolutional Neural Networks to better recognize objects of different scales in natural images.
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 26 REFERENCES
Deep machine learning provides state-of-the-art performance in image-based plant phenotyping
TLDR
Deep learning–based phenotyping is shown to have very good detection and localization accuracy in validation and testing image sets and to derive meaningful biological traits, which in turn can be used in quantitative trait loci discovery pipelines.
Deep Phenotyping: Deep Learning for Temporal Phenotype/Genotype Classification
TLDR
This paper exploits the power of deep CNNs for joint feature and classifier learning, within an automatic phenotyping scheme for genotype classification, and demonstrates that temporal information further improves the performance of the phenotype classification system.
Machine Learning for Plant Phenotyping Needs Image Processing.
Machine Learning for High-Throughput Stress Phenotyping in Plants.
Utilizing machine learning approaches to improve the prediction of leaf counts and individual leaf segmentation of rosette plant images
TLDR
An approach including image analysis methods for extraction of a comprehensive set of image features (based on the software IAP) to predict the number of leaves for rosette plants and a method to detect exact leaf borders for resolving inaccurate leaf segmentation in case of leaf overlaps are developed.
Learning to Count Leaves in Rosette Plants
TLDR
These findings suggest that it is possible to treat leaf counting as a regression problem, requiring as input only the total leaf count per training image, and relate image-based descriptors learned in an unsupervised fashion to leaf counts using a supervised regression model.
Metro Maps of Plant Disease Dynamics—Automated Mining of Differences Using Hyperspectral Images
TLDR
A cascade of data mining techniques for fast and reliable data-driven sketching of complex hyperspectral dynamics in plant science and plant phenotyping and uses the archetypal signatures to create structured summaries that are inspired by metro maps, i.e. schematic diagrams of public transport networks.
Computer vision cracks the leaf code
TLDR
It is demonstrated that a computer vision algorithm trained on several thousand images of diverse cleared leaves successfully learns leaf-architectural features, then categorizes novel specimens into natural botanical groups above the species level.
Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation
TLDR
This paper proposes a simple and scalable detection algorithm that improves mean average precision (mAP) by more than 30% relative to the previous best result on VOC 2012 -- achieving a mAP of 53.3%.
...
1
2
3
...