{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2024,7,27]],"date-time":"2024-07-27T07:51:40Z","timestamp":1722066700010},"reference-count":20,"publisher":"Oxford University Press (OUP)","issue":"4","license":[{"start":{"date-parts":[[2021,2,1]],"date-time":"2021-02-01T00:00:00Z","timestamp":1612137600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/journals\/pages\/open_access\/funder_policies\/chorus\/standard_publication_model"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2022,7,25]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Robustness is a significant constraint in machine learning models. The performance of the algorithms must not deteriorate when training and testing with slightly different data. Deep neural network models achieve awe-inspiring results in a wide range of applications of computer vision. Still, in the presence of noise or region occlusion, some models exhibit inaccurate performance even with data handled in training. Besides, some experiments suggest deep learning models sometimes use incorrect parts of the input information to perform inference. Active image augmentation (ADA) is an augmentation method that uses interpretability methods to augment the training data and improve its robustness to face the described problems. Although ADA presented interesting results, its original version only used the vanilla backpropagation interpretability to train the U-Net model. In this work, we propose an extensive experimental analysis of the interpretability method\u2019s impact on ADA. We use five interpretability methods: vanilla backpropagation, guided backpropagation, gradient-weighted class activation mapping (GradCam), guided GradCam and InputXGradient. The results show that all methods achieve similar performance at the ending of training, but when combining ADA with GradCam, the U-Net model presented an impressive fast convergence.<\/jats:p>","DOI":"10.1093\/jigpal\/jzab006","type":"journal-article","created":{"date-parts":[[2021,1,14]],"date-time":"2021-01-14T10:21:43Z","timestamp":1610619703000},"page":"611-621","source":"Crossref","is-referenced-by-count":2,"title":["On the Impact of Interpretability Methods in Active Image Augmentation Method"],"prefix":"10.1093","volume":"30","author":[{"given":"Fl\u00e1vio","family":"Arthur Oliveira Santos","sequence":"first","affiliation":[{"name":"Centro de Inform\u00e1tica , Universidade Federal de Pernambuco, Recife, 50670-901, Brasil"}]},{"given":"Cleber","family":"Zanchettin","sequence":"additional","affiliation":[{"name":"Centro de Inform\u00e1tica , Universidade Federal de Pernambuco, Recife, 50670-901, Brasil"}]},{"given":"Leonardo","family":"Nogueira Matos","sequence":"additional","affiliation":[{"name":"Departamento de Computa\u00e7\u00e3o , Universidade Federal de Sergipe, S\u00e3o Crist\u00f3v\u00e3o, 49100-000, Brasil"}]},{"given":"Paulo","family":"Novais","sequence":"additional","affiliation":[{"name":"Department of Informatics , University of Minho, Braga, 4710-057, Portugal"}]}],"member":"286","published-online":{"date-parts":[[2021,2,1]]},"reference":[{"key":"2022072013380964600_ref1","doi-asserted-by":"crossref","first-page":"297","DOI":"10.2307\/1932409","article-title":"Measures of the amount of ecologic association between species","volume":"26","author":"Dice","year":"1945","journal-title":"Ecology"},{"key":"2022072013380964600_ref2","article-title":"In defense of grid features for visual question answering","author":"Jiang","year":"2020","journal-title":"IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)"},{"key":"2022072013380964600_ref3","article-title":"Investigating the influence of noise and distractors on the interpretation of neural networks","author":"Kindermans","year":"2016"},{"key":"2022072013380964600_ref4","article-title":"Adam: A method for stochastic optimization","volume-title":"3rd International Conference on Learning Representations (ICLR)","author":"Kingma","year":"2015"},{"key":"2022072013380964600_ref5","article-title":"Pointrend: image segmentation as rendering","author":"Kirillov","year":"2020","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)"},{"key":"2022072013380964600_ref6","doi-asserted-by":"crossref","first-page":"436","DOI":"10.1038\/nature14539","article-title":"Deep learning","volume":"521","author":"LeCun","year":"2015","journal-title":"Nature"},{"key":"2022072013380964600_ref7","first-page":"1995","article-title":"Convolutional networks for images, speech, and time series","volume":"3361","author":"LeCun","year":"1995","journal-title":"The Handbook of Brain Theory and Neural Networks"},{"key":"2022072013380964600_ref8","first-page":"807","article-title":"Rectified linear units improve restricted boltzmann machines","volume-title":"Proceedings of the 27th International Conference on Machine Learning (ICML-10)","author":"Nair","year":"2010"},{"key":"2022072013380964600_ref9","doi-asserted-by":"crossref","first-page":"5966","DOI":"10.1038\/s41598-018-24304-3","article-title":"Spinal cord gray matter segmentation using deep dilated convolutions","volume":"8","author":"Perone","year":"2018","journal-title":"Scientific Reports"},{"key":"2022072013380964600_ref10","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-319-67558-9_38","article-title":"Grey matter segmentation in spinal cord mris via 3d convolutional encoder networks with shortcut connections","volume-title":"Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support, pp. 330\u2013337. Springer","author":"Porisky","year":"2017"},{"key":"2022072013380964600_ref11","doi-asserted-by":"crossref","first-page":"312","DOI":"10.1016\/j.neuroimage.2017.03.010","article-title":"Spinal cord grey matter segmentation challenge","volume":"152","author":"Prados","year":"2017","journal-title":"Neuroimage"},{"key":"2022072013380964600_ref12","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-319-24574-4_28","article-title":"U-net: convolutional networks for biomedical image segmentation","volume-title":"International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 234\u2013241. Springer","author":"Ronneberger","year":"2015"},{"key":"2022072013380964600_ref13","doi-asserted-by":"crossref","first-page":"310","DOI":"10.1007\/978-3-030-29859-3_27","article-title":"Active image data augmentation","volume-title":"Hybrid Artificial Intelligent Systems","author":"Santos","year":"2019"},{"key":"2022072013380964600_ref14","first-page":"618","article-title":"Grad-cam: visual explanations from deep networks via gradient-based localization","author":"Selvaraju","year":"2017","journal-title":"Proceedings of the IEEE International Conference on Computer Vision"},{"key":"2022072013380964600_ref15","article-title":"Deep inside convolutional networks: visualising image classification models and saliency maps","volume-title":"2nd International Conference on Learning Representations, {ICLR} 2014","author":"Simonyan","year":"2014"},{"key":"2022072013380964600_ref16","article-title":"Striving for simplicity: the all convolutional net","volume-title":"3rd International Conference on Learning Representations (ICLR)","author":"Springenberg","year":"2015"},{"key":"2022072013380964600_ref17","article-title":"Axiomatic attribution for deep networks","volume-title":"Proceedings of the 34th International Conference on Machine Learning (ICML)","author":"Sundararajan","year":"2017"},{"key":"2022072013380964600_ref18","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-319-10590-1_53","article-title":"Visualizing and understanding convolutional networks","volume-title":"European Conference on Computer Vision. pp. 818\u2013833. Springer","author":"Zeiler","year":"2014"},{"key":"2022072013380964600_ref19","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2010.5539957","volume-title":"Deconvolutional Networks","author":"Zeiler","year":"2010"},{"key":"2022072013380964600_ref20","first-page":"286","article-title":"Image super-resolution using very deep residual channel attention networks","author":"Zhang","year":"2018","journal-title":"Proceedings of the European Conference on Computer Vision (ECCV)"}],"container-title":["Logic Journal of the IGPL"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/jigpal\/article-pdf\/30\/4\/611\/45027460\/jzab006.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jigpal\/article-pdf\/30\/4\/611\/45027460\/jzab006.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,7,20]],"date-time":"2022-07-20T13:39:12Z","timestamp":1658324352000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jigpal\/article\/30\/4\/611\/6123345"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,2,1]]},"references-count":20,"journal-issue":{"issue":"4","published-online":{"date-parts":[[2021,2,1]]},"published-print":{"date-parts":[[2022,7,25]]}},"URL":"https:\/\/doi.org\/10.1093\/jigpal\/jzab006","relation":{},"ISSN":["1367-0751","1368-9894"],"issn-type":[{"value":"1367-0751","type":"print"},{"value":"1368-9894","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2022,7,19]]},"published":{"date-parts":[[2021,2,1]]}}}