Scalable learning for bridging the species gap in image-based plant phenotyping

Published in

Computer Vision and Image Understanding(2020)

External Links:

Generate Graph DownloadPDF

TL;DR

This paper investigates the use of synthetic data for leaf instance segmentation and presents UPGen: a Universal Plant Generator for bridging the species gap, which leverages domain randomisation to produce widely distributed data samples and models stochastic biological variation.

Authors

Peyman Moghadam

2 papers

Daniel Ward

1 papers

References67 items

Optimising realism of synthetic images using cycle generative adversarial networks for improved part segmentation

Seeing What a GAN Cannot Generate

Crop Phenomics: Current Status and Perspectives

Emergence of Object Segmentation in Perturbed Generative Models

Learn to synthesize and synthesize to learn

Scalable learning for bridging the species gap in image-based plant phenotyping

Published in

Computer Vision and Image Understanding(2020)

External Links:

Generate Graph DownloadPDF

TL;DR

Authors

Peyman Moghadam

2 papers

Daniel Ward

1 papers

References67 items

Optimising realism of synthetic images using cycle generative adversarial networks for improved part segmentation

Seeing What a GAN Cannot Generate

Crop Phenomics: Current Status and Perspectives

Emergence of Object Segmentation in Perturbed Generative Models

Learn to synthesize and synthesize to learn

Instance Segmentation of Biological Images Using Harmonic Embeddings

Learning to Generate Synthetic Data via Compositing

Plant Phenotyping: Past, Present, and Future

Data Augmentation for Leaf Segmentation and Counting Tasks in Rosette Plants

Latent Space Phenotyping: Automatic Image-Based Phenotyping for Treatment Studies

Sharing the Right Data Right: A Symbiosis with Machine Learning.

Can Adversarial Networks Hallucinate Occluded People With a Plausible Aspect?

Domain Randomization for Scene-Specific Car Detection and Pose Estimation

GAN Dissection: Visualizing and Understanding Generative Adversarial Networks

Data Augmentation using Conditional Generative Adversarial Networks for Leaf Counting in Arabidopsis Plants

Instance Segmentation by Deep Coloring

Modeling Visual Context is Key to Augmenting Object Detection Datasets

Deep Leaf Segmentation Using Synthetic Data

Constructing Unrestricted Adversarial Examples with Generative Models

Training Deep Networks with Synthetic Data: Bridging the Reality Gap by Domain Randomization

Adversarial Framework for Unsupervised Learning of Motion Dynamics in Videos

Improved Part Segmentation Performance by Optimising Realism of Synthetic Images using Cycle Generative Adversarial Networks

ST-GAN: Spatial Transformer Generative Adversarial Networks for Image Compositing

The use of plant models in deep learning: an application to leaf counting in rosette plants

Transfer Learning by Finetuning Pretrained CNNs Entirely with Synthetic Images

Recurrent Neural Networks for Semantic Instance Segmentation

On Pre-Trained Image Features and Synthetic Images for Deep Learning

weedNet: Dense Semantic Weed Classification Using Multispectral Images and MAV for Smart Farming

ARIGAN: Synthetic Arabidopsis Plants using Generative Adversarial Network

Leaf-GP: an open and automated software application for measuring growth phenotypes for arabidopsis and wheat

Sim4CV: A Photo-Realistic Simulator for Computer Vision Applications

Semantic Instance Segmentation with a Discriminative Loss Function

Cut, Paste and Learn: Surprisingly Easy Synthesis for Instance Detection

Revisiting Unreasonable Effectiveness of Data in Deep Learning Era

An Easy-to-Setup 3D Phenotyping Platform for KOMATSUNA Dataset

Morphometrics Reveals Complex and Heritable Apple Leaf Shapes

Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks

Domain randomization for transferring deep neural networks from simulation to the real world

Peduncle Detection of Sweet Pepper for Autonomous Crop Harvesting—Combined Color and 3-D Information

RenderGAN: Generating Realistic Labeled Data

Finely-grained annotated datasets for image-based plant phenotyping

The SYNTHIA Dataset: A Large Collection of Synthetic Images for Semantic Segmentation of Urban Scenes

End-to-End Instance Segmentation with Recurrent Attention

Leaf segmentation in plant phenotyping: a collation study

Recurrent Instance Segmentation

Deep-plant: Plant identification with convolutional neural networks

Image Analysis: The New Bottleneck in Plant Phenotyping [Applications Corner]

3-D Histogram-Based Segmentation and Leaf Detection for Rosette Plants

ImageNet Large Scale Visual Recognition Challenge

Annotated Image Datasets of Rosette Plants

The Role of Context for Object Detection and Semantic Segmentation in the Wild

Microsoft COCO: Common Objects in Context

Multi-leaf tracking from fluorescence plant videos

Vision meets robotics: The KITTI dataset

SUN database: Large-scale scene recognition from abbey to zoo

The Unreasonable Effectiveness of Data

Using Simulation and Domain Adaptation to Improve Efficiency of Deep Robotic Grasping

Data synthesis methods for semantic segmentation in agriculture: A Capsicum annuum dataset

Synthetic arabidopsis

Plant Disease Detection Using Hyperspectral Imaging

Openimages: A public dataset for large-scale multi-label and multi-class image classification

Aberystwyth leaf evaluation dataset

Computer vision and pattern recognition (CVPR), IEEE

Trends in Plant Science

ART AND SCIENCE OF LIFE: DESIGNING AND GROWING VIRTUAL PLANTS WITH L-SYSTEMS

Example synthetic plant images generated by the UPGen pipeline. Five column pairs present example RGB images (left column) and the corresponding ground truth per-pixel leaf instance annotations

Field of Study

Computer Science

Journal Information

Name

ArXiv

Volume

abs/2005.00687

Venue Information

Name

Computer Vision and Image Understanding

Type

journal

URL

http://www.elsevier.com/wps/find/journaldescription.cws_home/622809/description#description

Alternate Names

Comput Vis Image Underst

TL;DR

Authors

References67 items

Optimising realism of synthetic images using cycle generative adversarial networks for improved part segmentation

Seeing What a GAN Cannot Generate

Crop Phenomics: Current Status and Perspectives

Emergence of Object Segmentation in Perturbed Generative Models

Learn to synthesize and synthesize to learn

TL;DR

Authors

References67 items

Optimising realism of synthetic images using cycle generative adversarial networks for improved part segmentation

Seeing What a GAN Cannot Generate

Crop Phenomics: Current Status and Perspectives

Emergence of Object Segmentation in Perturbed Generative Models

Learn to synthesize and synthesize to learn

Instance Segmentation of Biological Images Using Harmonic Embeddings

Learning to Generate Synthetic Data via Compositing

Plant Phenotyping: Past, Present, and Future

Data Augmentation for Leaf Segmentation and Counting Tasks in Rosette Plants

Latent Space Phenotyping: Automatic Image-Based Phenotyping for Treatment Studies

Sharing the Right Data Right: A Symbiosis with Machine Learning.

Can Adversarial Networks Hallucinate Occluded People With a Plausible Aspect?

Domain Randomization for Scene-Specific Car Detection and Pose Estimation

GAN Dissection: Visualizing and Understanding Generative Adversarial Networks

Data Augmentation using Conditional Generative Adversarial Networks for Leaf Counting in Arabidopsis Plants

Instance Segmentation by Deep Coloring

Modeling Visual Context is Key to Augmenting Object Detection Datasets

Deep Leaf Segmentation Using Synthetic Data

Constructing Unrestricted Adversarial Examples with Generative Models

Training Deep Networks with Synthetic Data: Bridging the Reality Gap by Domain Randomization

Adversarial Framework for Unsupervised Learning of Motion Dynamics in Videos

Improved Part Segmentation Performance by Optimising Realism of Synthetic Images using Cycle Generative Adversarial Networks

ST-GAN: Spatial Transformer Generative Adversarial Networks for Image Compositing

The use of plant models in deep learning: an application to leaf counting in rosette plants

Transfer Learning by Finetuning Pretrained CNNs Entirely with Synthetic Images

Recurrent Neural Networks for Semantic Instance Segmentation

On Pre-Trained Image Features and Synthetic Images for Deep Learning

weedNet: Dense Semantic Weed Classification Using Multispectral Images and MAV for Smart Farming

ARIGAN: Synthetic Arabidopsis Plants using Generative Adversarial Network

Leaf-GP: an open and automated software application for measuring growth phenotypes for arabidopsis and wheat

Sim4CV: A Photo-Realistic Simulator for Computer Vision Applications

Semantic Instance Segmentation with a Discriminative Loss Function

Cut, Paste and Learn: Surprisingly Easy Synthesis for Instance Detection

Revisiting Unreasonable Effectiveness of Data in Deep Learning Era

An Easy-to-Setup 3D Phenotyping Platform for KOMATSUNA Dataset

Morphometrics Reveals Complex and Heritable Apple Leaf Shapes

Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks

Domain randomization for transferring deep neural networks from simulation to the real world

Mask R-CNN

Peduncle Detection of Sweet Pepper for Autonomous Crop Harvesting—Combined Color and 3-D Information

RenderGAN: Generating Realistic Labeled Data

Finely-grained annotated datasets for image-based plant phenotyping

The SYNTHIA Dataset: A Large Collection of Synthetic Images for Semantic Segmentation of Urban Scenes

End-to-End Instance Segmentation with Recurrent Attention

Leaf segmentation in plant phenotyping: a collation study

Recurrent Instance Segmentation

Deep-plant: Plant identification with convolutional neural networks

Image Analysis: The New Bottleneck in Plant Phenotyping [Applications Corner]

3-D Histogram-Based Segmentation and Leaf Detection for Rosette Plants

ImageNet Large Scale Visual Recognition Challenge

Annotated Image Datasets of Rosette Plants

The Role of Context for Object Detection and Semantic Segmentation in the Wild

Microsoft COCO: Common Objects in Context

Multi-leaf tracking from fluorescence plant videos

Vision meets robotics: The KITTI dataset

SUN database: Large-scale scene recognition from abbey to zoo

The Unreasonable Effectiveness of Data

Using Simulation and Domain Adaptation to Improve Efficiency of Deep Robotic Grasping

Data synthesis methods for semantic segmentation in agriculture: A Capsicum annuum dataset

Synthetic arabidopsis

Plant Disease Detection Using Hyperspectral Imaging

Openimages: A public dataset for large-scale multi-label and multi-class image classification

Aberystwyth leaf evaluation dataset

Computer vision and pattern recognition (CVPR), IEEE

Trends in Plant Science

ART AND SCIENCE OF LIFE: DESIGNING AND GROWING VIRTUAL PLANTS WITH L-SYSTEMS

Example synthetic plant images generated by the UPGen pipeline. Five column pairs present example RGB images (left column) and the corresponding ground truth per-pixel leaf instance annotations

Field of Study

Journal Information