ARIGAN: Synthetic Arabidopsis Plants Using Generative Adversarial Network. Giuffrida, M. V., Scharr, H., & Tsaftaris, S. A. In 2017 IEEE International Conference on Computer Vision Workshops (ICCVW), pages 2064–2071, October, 2017. doi abstract bibtex In recent years, there has been an increasing interest in image-based plant phenotyping, applying state-of-the-art machine learning approaches to tackle challenging problems, such as leaf segmentation (a multi-instance problem) and counting. Most of these algorithms need labelled data to learn a model for the task at hand. Despite the recent release of a few plant phenotyping datasets, large annotated plant image datasets for the purpose of training deep learning algorithms are lacking. One common approach to alleviate the lack of training data is dataset augmentation. Herein, we propose an alternative solution to dataset aug-mentation for plant phenotyping, creating artificial images of plants using generative neural networks. We propose the Arabidopsis Rosette Image Generator (through) Adversarial Network: a deep convolutional network that is able to generate synthetic rosette-shaped plants, inspired by DC-GAN (a recent adversarial network model using convolutional layers). Specifically, we trained the network using A1, A2, and A4 of the CVPPP 2017 LCC dataset, containing Arabidopsis Thaliana plants. We show that our model is able to generate realistic 128 x 128 colour images of plants. We train our network conditioning on leaf count, such that it is possible to generate plants with a given number of leaves suitable, among others, for training regression based models. We propose a new Ax dataset of artificial plants images, obtained by our ARIGAN. We evaluate this new dataset using a state-of-the-art leaf counting algorithm, showing that the testing error is reduced when Ax is used as part of the training data.
@inproceedings{giuffrida_arigan:_2017,
title = {{ARIGAN}: {Synthetic} {Arabidopsis} {Plants} {Using} {Generative} {Adversarial} {Network}},
shorttitle = {{ARIGAN}},
doi = {10.1109/ICCVW.2017.242},
abstract = {In recent years, there has been an increasing interest in image-based plant phenotyping, applying state-of-the-art machine learning approaches to tackle challenging problems, such as leaf segmentation (a multi-instance problem) and counting. Most of these algorithms need labelled data to learn a model for the task at hand. Despite the recent release of a few plant phenotyping datasets, large annotated plant image datasets for the purpose of training deep learning algorithms are lacking. One common approach to alleviate the lack of training data is dataset augmentation. Herein, we propose an alternative solution to dataset aug-mentation for plant phenotyping, creating artificial images of plants using generative neural networks. We propose the Arabidopsis Rosette Image Generator (through) Adversarial Network: a deep convolutional network that is able to generate synthetic rosette-shaped plants, inspired by DC-GAN (a recent adversarial network model using convolutional layers). Specifically, we trained the network using A1, A2, and A4 of the CVPPP 2017 LCC dataset, containing Arabidopsis Thaliana plants. We show that our model is able to generate realistic 128 x 128 colour images of plants. We train our network conditioning on leaf count, such that it is possible to generate plants with a given number of leaves suitable, among others, for training regression based models. We propose a new Ax dataset of artificial plants images, obtained by our ARIGAN. We evaluate this new dataset using a state-of-the-art leaf counting algorithm, showing that the testing error is reduced when Ax is used as part of the training data.},
booktitle = {2017 {IEEE} {International} {Conference} on {Computer} {Vision} {Workshops} ({ICCVW})},
author = {Giuffrida, M. V. and Scharr, H. and Tsaftaris, S. A.},
month = oct,
year = {2017},
keywords = {ARIGAN, Arabidopsis Rosette Image Generator, Arabidopsis Thaliana plants, Ax dataset, CVPPP 2017 LCC dataset, Computational modeling, Computer vision, Data models, Gallium nitride, Generators, Neural networks, Training, adversarial network model, annotated plant image datasets, artificial images, artificial plants images, biology computing, botany, dataset augmentation, deep convolutional network, deep learning algorithms, generative adversarial network, generative neural networks, leaf count, leaf segmentation, learning (artificial intelligence), network conditioning, neural nets, plant phenotyping datasets, realistic colour images, realistic images, regression analysis, regression based models, synthetic Arabidopsis plants, synthetic rosette-shaped plants, training data},
pages = {2064--2071}
}
Downloads: 0
{"_id":"7r4zrRvW56xZWAt4a","bibbaseid":"giuffrida-scharr-tsaftaris-arigansyntheticarabidopsisplantsusinggenerativeadversarialnetwork-2017","downloads":0,"creationDate":"2019-03-27T01:00:55.583Z","title":"ARIGAN: Synthetic Arabidopsis Plants Using Generative Adversarial Network","author_short":["Giuffrida, M. V.","Scharr, H.","Tsaftaris, S. A."],"year":2017,"bibtype":"inproceedings","biburl":"https://api.zotero.org/groups/2196137/items?key=9sAVQ9xrV4yCcI9wtDczc9Dg&format=bibtex&limit=100","bibdata":{"bibtype":"inproceedings","type":"inproceedings","title":"ARIGAN: Synthetic Arabidopsis Plants Using Generative Adversarial Network","shorttitle":"ARIGAN","doi":"10.1109/ICCVW.2017.242","abstract":"In recent years, there has been an increasing interest in image-based plant phenotyping, applying state-of-the-art machine learning approaches to tackle challenging problems, such as leaf segmentation (a multi-instance problem) and counting. Most of these algorithms need labelled data to learn a model for the task at hand. Despite the recent release of a few plant phenotyping datasets, large annotated plant image datasets for the purpose of training deep learning algorithms are lacking. One common approach to alleviate the lack of training data is dataset augmentation. Herein, we propose an alternative solution to dataset aug-mentation for plant phenotyping, creating artificial images of plants using generative neural networks. We propose the Arabidopsis Rosette Image Generator (through) Adversarial Network: a deep convolutional network that is able to generate synthetic rosette-shaped plants, inspired by DC-GAN (a recent adversarial network model using convolutional layers). Specifically, we trained the network using A1, A2, and A4 of the CVPPP 2017 LCC dataset, containing Arabidopsis Thaliana plants. We show that our model is able to generate realistic 128 x 128 colour images of plants. We train our network conditioning on leaf count, such that it is possible to generate plants with a given number of leaves suitable, among others, for training regression based models. We propose a new Ax dataset of artificial plants images, obtained by our ARIGAN. We evaluate this new dataset using a state-of-the-art leaf counting algorithm, showing that the testing error is reduced when Ax is used as part of the training data.","booktitle":"2017 IEEE International Conference on Computer Vision Workshops (ICCVW)","author":[{"propositions":[],"lastnames":["Giuffrida"],"firstnames":["M.","V."],"suffixes":[]},{"propositions":[],"lastnames":["Scharr"],"firstnames":["H."],"suffixes":[]},{"propositions":[],"lastnames":["Tsaftaris"],"firstnames":["S.","A."],"suffixes":[]}],"month":"October","year":"2017","keywords":"ARIGAN, Arabidopsis Rosette Image Generator, Arabidopsis Thaliana plants, Ax dataset, CVPPP 2017 LCC dataset, Computational modeling, Computer vision, Data models, Gallium nitride, Generators, Neural networks, Training, adversarial network model, annotated plant image datasets, artificial images, artificial plants images, biology computing, botany, dataset augmentation, deep convolutional network, deep learning algorithms, generative adversarial network, generative neural networks, leaf count, leaf segmentation, learning (artificial intelligence), network conditioning, neural nets, plant phenotyping datasets, realistic colour images, realistic images, regression analysis, regression based models, synthetic Arabidopsis plants, synthetic rosette-shaped plants, training data","pages":"2064–2071","bibtex":"@inproceedings{giuffrida_arigan:_2017,\n\ttitle = {{ARIGAN}: {Synthetic} {Arabidopsis} {Plants} {Using} {Generative} {Adversarial} {Network}},\n\tshorttitle = {{ARIGAN}},\n\tdoi = {10.1109/ICCVW.2017.242},\n\tabstract = {In recent years, there has been an increasing interest in image-based plant phenotyping, applying state-of-the-art machine learning approaches to tackle challenging problems, such as leaf segmentation (a multi-instance problem) and counting. Most of these algorithms need labelled data to learn a model for the task at hand. Despite the recent release of a few plant phenotyping datasets, large annotated plant image datasets for the purpose of training deep learning algorithms are lacking. One common approach to alleviate the lack of training data is dataset augmentation. Herein, we propose an alternative solution to dataset aug-mentation for plant phenotyping, creating artificial images of plants using generative neural networks. We propose the Arabidopsis Rosette Image Generator (through) Adversarial Network: a deep convolutional network that is able to generate synthetic rosette-shaped plants, inspired by DC-GAN (a recent adversarial network model using convolutional layers). Specifically, we trained the network using A1, A2, and A4 of the CVPPP 2017 LCC dataset, containing Arabidopsis Thaliana plants. We show that our model is able to generate realistic 128 x 128 colour images of plants. We train our network conditioning on leaf count, such that it is possible to generate plants with a given number of leaves suitable, among others, for training regression based models. We propose a new Ax dataset of artificial plants images, obtained by our ARIGAN. We evaluate this new dataset using a state-of-the-art leaf counting algorithm, showing that the testing error is reduced when Ax is used as part of the training data.},\n\tbooktitle = {2017 {IEEE} {International} {Conference} on {Computer} {Vision} {Workshops} ({ICCVW})},\n\tauthor = {Giuffrida, M. V. and Scharr, H. and Tsaftaris, S. A.},\n\tmonth = oct,\n\tyear = {2017},\n\tkeywords = {ARIGAN, Arabidopsis Rosette Image Generator, Arabidopsis Thaliana plants, Ax dataset, CVPPP 2017 LCC dataset, Computational modeling, Computer vision, Data models, Gallium nitride, Generators, Neural networks, Training, adversarial network model, annotated plant image datasets, artificial images, artificial plants images, biology computing, botany, dataset augmentation, deep convolutional network, deep learning algorithms, generative adversarial network, generative neural networks, leaf count, leaf segmentation, learning (artificial intelligence), network conditioning, neural nets, plant phenotyping datasets, realistic colour images, realistic images, regression analysis, regression based models, synthetic Arabidopsis plants, synthetic rosette-shaped plants, training data},\n\tpages = {2064--2071}\n}\n\n","author_short":["Giuffrida, M. V.","Scharr, H.","Tsaftaris, S. A."],"key":"giuffrida_arigan:_2017","id":"giuffrida_arigan:_2017","bibbaseid":"giuffrida-scharr-tsaftaris-arigansyntheticarabidopsisplantsusinggenerativeadversarialnetwork-2017","role":"author","urls":{},"keyword":["ARIGAN","Arabidopsis Rosette Image Generator","Arabidopsis Thaliana plants","Ax dataset","CVPPP 2017 LCC dataset","Computational modeling","Computer vision","Data models","Gallium nitride","Generators","Neural networks","Training","adversarial network model","annotated plant image datasets","artificial images","artificial plants images","biology computing","botany","dataset augmentation","deep convolutional network","deep learning algorithms","generative adversarial network","generative neural networks","leaf count","leaf segmentation","learning (artificial intelligence)","network conditioning","neural nets","plant phenotyping datasets","realistic colour images","realistic images","regression analysis","regression based models","synthetic Arabidopsis plants","synthetic rosette-shaped plants","training data"],"downloads":0,"html":""},"search_terms":["arigan","synthetic","arabidopsis","plants","using","generative","adversarial","network","giuffrida","scharr","tsaftaris"],"keywords":["arigan","arabidopsis rosette image generator","arabidopsis thaliana plants","ax dataset","cvppp 2017 lcc dataset","computational modeling","computer vision","data models","gallium nitride","generators","neural networks","training","adversarial network model","annotated plant image datasets","artificial images","artificial plants images","biology computing","botany","dataset augmentation","deep convolutional network","deep learning algorithms","generative adversarial network","generative neural networks","leaf count","leaf segmentation","learning (artificial intelligence)","network conditioning","neural nets","plant phenotyping datasets","realistic colour images","realistic images","regression analysis","regression based models","synthetic arabidopsis plants","synthetic rosette-shaped plants","training data"],"authorIDs":[],"dataSources":["M8pbA37dQECNJ7i5D"]}