Self-Supervised Out-of-Distribution Detection in Brain CT Scans

Self-Supervised Out-of-Distribution Detection in Brain CT Scans. Venkatakrishnan, A., R., Kim, S., T., Eisawy, R., Pfister, F., & Navab, N. arXiv:2011.05428 [cs, eess], 11, 2020.

Paper

Self-Supervised Out-of-Distribution Detection in Brain CT Scans [link]

Website abstract bibtex

Medical imaging data suffers from the limited availability of annotation because annotating 3D medical data is a time-consuming and expensive task. Moreover, even if the annotation is available, supervised learning-based approaches suffer highly imbalanced data. Most of the scans during the screening are from normal subjects, but there are also large variations in abnormal cases. To address these issues, recently, unsupervised deep anomaly detection methods that train the model on large-sized normal scans and detect abnormal scans by calculating reconstruction error have been reported. In this paper, we propose a novel self-supervised learning technique for anomaly detection. Our architecture largely consists of two parts: 1) Reconstruction and 2) predicting geometric transformations. By training the network to predict geometric transformations, the model could learn better image features and distribution of normal scans. In the test time, the geometric transformation predictor can assign the anomaly score by calculating the error between geometric transformation and prediction. Moreover, we further use self-supervised learning with context restoration for pretraining our model. By comparative experiments on clinical brain CT scans, the effectiveness of the proposed method has been verified.

@article{
 title = {Self-Supervised Out-of-Distribution Detection in Brain CT Scans},
 type = {article},
 year = {2020},
 keywords = {Computer Science - Computer Vision and Pattern Rec,Electrical Engineering and Systems Science - Imag},
 websites = {http://arxiv.org/abs/2011.05428},
 month = {11},
 id = {82bdd8b6-3f15-3a50-b08e-e485c074bcd6},
 created = {2022-03-28T09:45:00.683Z},
 accessed = {2022-03-26},
 file_attached = {true},
 profile_id = {235249c2-3ed4-314a-b309-b1ea0330f5d9},
 group_id = {1ff583c0-be37-34fa-9c04-73c69437d354},
 last_modified = {2022-03-29T08:01:10.407Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 citation_key = {venkatakrishnanSelfSupervisedOutofDistributionDetection2020},
 source_type = {article},
 notes = {arXiv: 2011.05428},
 private_publication = {false},
 abstract = {Medical imaging data suffers from the limited availability of annotation because annotating 3D medical data is a time-consuming and expensive task. Moreover, even if the annotation is available, supervised learning-based approaches suffer highly imbalanced data. Most of the scans during the screening are from normal subjects, but there are also large variations in abnormal cases. To address these issues, recently, unsupervised deep anomaly detection methods that train the model on large-sized normal scans and detect abnormal scans by calculating reconstruction error have been reported. In this paper, we propose a novel self-supervised learning technique for anomaly detection. Our architecture largely consists of two parts: 1) Reconstruction and 2) predicting geometric transformations. By training the network to predict geometric transformations, the model could learn better image features and distribution of normal scans. In the test time, the geometric transformation predictor can assign the anomaly score by calculating the error between geometric transformation and prediction. Moreover, we further use self-supervised learning with context restoration for pretraining our model. By comparative experiments on clinical brain CT scans, the effectiveness of the proposed method has been verified.},
 bibtype = {article},
 author = {Venkatakrishnan, Abinav Ravi and Kim, Seong Tae and Eisawy, Rami and Pfister, Franz and Navab, Nassir},
 journal = {arXiv:2011.05428 [cs, eess]}
}

Downloads: 0

{"_id":"YPivSwjYuNS999xig","bibbaseid":"venkatakrishnan-kim-eisawy-pfister-navab-selfsupervisedoutofdistributiondetectioninbrainctscans-2020","author_short":["Venkatakrishnan, A., R.","Kim, S., T.","Eisawy, R.","Pfister, F.","Navab, N."],"bibdata":{"title":"Self-Supervised Out-of-Distribution Detection in Brain CT Scans","type":"article","year":"2020","keywords":"Computer Science - Computer Vision and Pattern Rec,Electrical Engineering and Systems Science - Imag","websites":"http://arxiv.org/abs/2011.05428","month":"11","id":"82bdd8b6-3f15-3a50-b08e-e485c074bcd6","created":"2022-03-28T09:45:00.683Z","accessed":"2022-03-26","file_attached":"true","profile_id":"235249c2-3ed4-314a-b309-b1ea0330f5d9","group_id":"1ff583c0-be37-34fa-9c04-73c69437d354","last_modified":"2022-03-29T08:01:10.407Z","read":false,"starred":false,"authored":false,"confirmed":"true","hidden":false,"citation_key":"venkatakrishnanSelfSupervisedOutofDistributionDetection2020","source_type":"article","notes":"arXiv: 2011.05428","private_publication":false,"abstract":"Medical imaging data suffers from the limited availability of annotation because annotating 3D medical data is a time-consuming and expensive task. Moreover, even if the annotation is available, supervised learning-based approaches suffer highly imbalanced data. Most of the scans during the screening are from normal subjects, but there are also large variations in abnormal cases. To address these issues, recently, unsupervised deep anomaly detection methods that train the model on large-sized normal scans and detect abnormal scans by calculating reconstruction error have been reported. In this paper, we propose a novel self-supervised learning technique for anomaly detection. Our architecture largely consists of two parts: 1) Reconstruction and 2) predicting geometric transformations. By training the network to predict geometric transformations, the model could learn better image features and distribution of normal scans. In the test time, the geometric transformation predictor can assign the anomaly score by calculating the error between geometric transformation and prediction. Moreover, we further use self-supervised learning with context restoration for pretraining our model. By comparative experiments on clinical brain CT scans, the effectiveness of the proposed method has been verified.","bibtype":"article","author":"Venkatakrishnan, Abinav Ravi and Kim, Seong Tae and Eisawy, Rami and Pfister, Franz and Navab, Nassir","journal":"arXiv:2011.05428 [cs, eess]","bibtex":"@article{\n title = {Self-Supervised Out-of-Distribution Detection in Brain CT Scans},\n type = {article},\n year = {2020},\n keywords = {Computer Science - Computer Vision and Pattern Rec,Electrical Engineering and Systems Science - Imag},\n websites = {http://arxiv.org/abs/2011.05428},\n month = {11},\n id = {82bdd8b6-3f15-3a50-b08e-e485c074bcd6},\n created = {2022-03-28T09:45:00.683Z},\n accessed = {2022-03-26},\n file_attached = {true},\n profile_id = {235249c2-3ed4-314a-b309-b1ea0330f5d9},\n group_id = {1ff583c0-be37-34fa-9c04-73c69437d354},\n last_modified = {2022-03-29T08:01:10.407Z},\n read = {false},\n starred = {false},\n authored = {false},\n confirmed = {true},\n hidden = {false},\n citation_key = {venkatakrishnanSelfSupervisedOutofDistributionDetection2020},\n source_type = {article},\n notes = {arXiv: 2011.05428},\n private_publication = {false},\n abstract = {Medical imaging data suffers from the limited availability of annotation because annotating 3D medical data is a time-consuming and expensive task. Moreover, even if the annotation is available, supervised learning-based approaches suffer highly imbalanced data. Most of the scans during the screening are from normal subjects, but there are also large variations in abnormal cases. To address these issues, recently, unsupervised deep anomaly detection methods that train the model on large-sized normal scans and detect abnormal scans by calculating reconstruction error have been reported. In this paper, we propose a novel self-supervised learning technique for anomaly detection. Our architecture largely consists of two parts: 1) Reconstruction and 2) predicting geometric transformations. By training the network to predict geometric transformations, the model could learn better image features and distribution of normal scans. In the test time, the geometric transformation predictor can assign the anomaly score by calculating the error between geometric transformation and prediction. Moreover, we further use self-supervised learning with context restoration for pretraining our model. By comparative experiments on clinical brain CT scans, the effectiveness of the proposed method has been verified.},\n bibtype = {article},\n author = {Venkatakrishnan, Abinav Ravi and Kim, Seong Tae and Eisawy, Rami and Pfister, Franz and Navab, Nassir},\n journal = {arXiv:2011.05428 [cs, eess]}\n}","author_short":["Venkatakrishnan, A., R.","Kim, S., T.","Eisawy, R.","Pfister, F.","Navab, N."],"urls":{"Paper":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c/file/35abb50a-9aee-7b73-6209-3ce75da99e92/Venkatakrishnan_et_al___2020___Self_Supervised_Out_of_Distribution_Detection_in_B.pdf.pdf","Website":"http://arxiv.org/abs/2011.05428"},"biburl":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c","bibbaseid":"venkatakrishnan-kim-eisawy-pfister-navab-selfsupervisedoutofdistributiondetectioninbrainctscans-2020","role":"author","keyword":["Computer Science - Computer Vision and Pattern Rec","Electrical Engineering and Systems Science - Imag"],"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c","dataSources":["PGS2HXygFqa56tPwp","ya2CyA73rpZseyrZ8","2252seNhipfTmjEBQ"],"keywords":["computer science - computer vision and pattern rec","electrical engineering and systems science - imag"],"search_terms":["self","supervised","out","distribution","detection","brain","scans","venkatakrishnan","kim","eisawy","pfister","navab"],"title":"Self-Supervised Out-of-Distribution Detection in Brain CT Scans","year":2020}