{"id":{"repo_id":"mcmaster","oai_identifier":"oai:macsphere.mcmaster.ca:11375/29573"},"canonical_url":"https://search.ndltd.org/etd/mcmaster/oai:macsphere.mcmaster.ca:11375/29573","repository":{"repo_id":"mcmaster","name":"McMaster University","base_url":"https://macsphere.mcmaster.ca/server/oai/request"},"display":{"title":"Two-way Multi-input Generative Neural Network for Anomaly Event Detection and Localization","abstract":"Anomaly event detection has become increasingly important and is of great significance for real-time monitoring systems. However, developing a reliable anomaly detection and localization model still requires overcoming many challenging problems considering the ambiguity in the definition of an abnormal event and the lack of ground truth datasets for training. In this thesis, we propose a Two-way Multi-input Generative Neural Network (TMGNN), which is an unsupervised anomaly events detection and localization method based on Generative Adversarial Network (GAN). TMGNN is composed of two neural networks, an appearance generation neural network and a motion generation neural network. These two networks are trained on normal frames and their corresponding motion and mosaic frames respectively. In the testing steps, the trained model cannot properly reconstruct the anomalous objects since the network is trained only on normal frames and has not learned patterns of anomalous cases. With the help of our new patch-based evaluation method, we utilize the reconstruction error to detect and localize possible anomalous objects. Our experiments show that on the UCSD Pedestrain2 dataset, our approach achieves 96.5% Area Under Curve (AUC) and 94.1% AUC for the frame-level and pixel-level criteria, respectively, reaching the best classification results compared to other traditional and deep learning methods.","abstract_html":"Anomaly event detection has become increasingly important and is of great significance for real-time monitoring systems. However, developing a reliable anomaly detection and localization model still requires overcoming many challenging problems considering the ambiguity in the definition of an abnormal event and the lack of ground truth datasets for training. In this thesis, we propose a Two-way Multi-input Generative Neural Network (TMGNN), which is an unsupervised anomaly events detection and localization method based on Generative Adversarial Network (GAN). TMGNN is composed of two neural networks, an appearance generation neural network and a motion generation neural network. These two networks are trained on normal frames and their corresponding motion and mosaic frames respectively. In the testing steps, the trained model cannot properly reconstruct the anomalous objects since the network is trained only on normal frames and has not learned patterns of anomalous cases. With the help of our new patch-based evaluation method, we utilize the reconstruction error to detect and localize possible anomalous objects. Our experiments show that on the UCSD Pedestrain2 dataset, our approach achieves 96.5% Area Under Curve (AUC) and 94.1% AUC for the frame-level and pixel-level criteria, respectively, reaching the best classification results compared to other traditional and deep learning methods.","abstract_has_math":false,"creators":["Yang, Mingchen"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":"Electrical and Computer Engineering","school":null,"contributors":[],"advisors":["Shirani, Shahram"],"committee_chairs":[],"committee_members":[],"year":2022,"date_issued":"2022","date_published":"2022","updated_at":"2026-11-19T05:56:56Z","subjects":["Anomaly detection and localization","Generative adversarial network","Two-way multi-input generative neural network"],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://doi.org/10.71548/23348"],"render_values":[{"text":"https://doi.org/10.71548/23348","href":"https://doi.org/10.71548/23348","code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/11375/29573","outbound_label":"Handle","outbound_source":"dc:identifier.uri","record_url":"http://hdl.handle.net/11375/29573","record_label":"Repository record","record_kind":"handle","record_source":"dc:identifier.uri","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://macsphere.mcmaster.ca/server/oai/request?verb=GetRecord&metadataPrefix=dim&identifier=oai%3Amacsphere.mcmaster.ca%3A11375%2F29573","prefix":"dim"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Shirani, Shahram"]},{"key":"dc:contributor.department","label":"Department","values":["Electrical and Computer Engineering"]},{"key":"dc:creator","label":"Author","values":["Yang, Mingchen"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2024-03-07T19:23:14Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2024-03-07T19:23:14Z"]},{"key":"dc:date.issued","label":"Date","values":["2022"]},{"key":"dc:type","label":"Dc Type","values":["Thesis","Image","Software","Video"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Anomaly detection and localization","Generative adversarial network","Two-way multi-input generative neural network"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["http://hdl.handle.net/11375/29573","https://doi.org/10.71548/23348"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Anomaly event detection has become increasingly important and is of great significance for real-time monitoring systems. However, developing a reliable anomaly detection and localization model still requires overcoming many challenging problems considering the ambiguity in the definition of an abnormal event and the lack of ground truth datasets for training. In this thesis, we propose a Two-way Multi-input Generative Neural Network (TMGNN), which is an unsupervised anomaly events detection and localization method based on Generative Adversarial Network (GAN). TMGNN is composed of two neural networks, an appearance generation neural network and a motion generation neural network. These two networks are trained on normal frames and their corresponding motion and mosaic frames respectively. In the testing steps, the trained model cannot properly reconstruct the anomalous objects since the network is trained only on normal frames and has not learned patterns of anomalous cases. With the help of our new patch-based evaluation method, we utilize the reconstruction error to detect and localize possible anomalous objects. Our experiments show that on the UCSD Pedestrain2 dataset, our approach achieves 96.5% Area Under Curve (AUC) and 94.1% AUC for the frame-level and pixel-level criteria, respectively, reaching the best classification results compared to other traditional and deep learning methods."]},{"key":"dc:description.abstractgeneral","label":"General Abstract","values":["Recently, abnormal event detection has attracted increasing attention in the field of surveillance video. However, it is still a big challenge to build an automatic and reliable abnormal event detection system to review a surveillance video containing hundreds of frames and mask the frames with abnormal objects or events. In this thesis, we build a model and teach it to memorize the structure of normal frames. Then the model is able to tell which frames are normal. Any other frames that appear in the surveillance video will be classified as abnormal frames. Moreover, we design a new method to evaluate the performance of our model and compare it with other models’ results."]},{"key":"dc:description.degree","label":"Dc Description Degree","values":["Master of Applied Science (MASc)"]},{"key":"dc:title","label":"Title","values":["Two-way Multi-input Generative Neural Network for Anomaly Event Detection and Localization"]},{"key":"others:access_status","label":"Others Access Status","values":["open.access"]}]}],"canonical_facts":{"dc:contributor.advisor":["Shirani, Shahram"],"dc:contributor.department":["Electrical and Computer Engineering"],"dc:creator":["Yang, Mingchen"],"dc:date.accessioned":["2024-03-07T19:23:14Z"],"dc:date.available":["2024-03-07T19:23:14Z"],"dc:date.issued":["2022"],"dc:description.abstract":["Anomaly event detection has become increasingly important and is of great significance for real-time monitoring systems. However, developing a reliable anomaly detection and localization model still requires overcoming many challenging problems considering the ambiguity in the definition of an abnormal event and the lack of ground truth datasets for training. In this thesis, we propose a Two-way Multi-input Generative Neural Network (TMGNN), which is an unsupervised anomaly events detection and localization method based on Generative Adversarial Network (GAN). TMGNN is composed of two neural networks, an appearance generation neural network and a motion generation neural network. These two networks are trained on normal frames and their corresponding motion and mosaic frames respectively. In the testing steps, the trained model cannot properly reconstruct the anomalous objects since the network is trained only on normal frames and has not learned patterns of anomalous cases. With the help of our new patch-based evaluation method, we utilize the reconstruction error to detect and localize possible anomalous objects. Our experiments show that on the UCSD Pedestrain2 dataset, our approach achieves 96.5% Area Under Curve (AUC) and 94.1% AUC for the frame-level and pixel-level criteria, respectively, reaching the best classification results compared to other traditional and deep learning methods."],"dc:description.abstractgeneral":["Recently, abnormal event detection has attracted increasing attention in the field of surveillance video. However, it is still a big challenge to build an automatic and reliable abnormal event detection system to review a surveillance video containing hundreds of frames and mask the frames with abnormal objects or events. In this thesis, we build a model and teach it to memorize the structure of normal frames. Then the model is able to tell which frames are normal. Any other frames that appear in the surveillance video will be classified as abnormal frames. Moreover, we design a new method to evaluate the performance of our model and compare it with other models’ results."],"dc:description.degree":["Master of Applied Science (MASc)"],"dc:identifier.uri":["http://hdl.handle.net/11375/29573","https://doi.org/10.71548/23348"],"dc:language.iso":["en"],"dc:subject":["Anomaly detection and localization","Generative adversarial network","Two-way multi-input generative neural network"],"dc:title":["Two-way Multi-input Generative Neural Network for Anomaly Event Detection and Localization"],"dc:type":["Thesis","Image","Software","Video"],"others:access_status":["open.access"]},"updated_at":"2026-11-19T05:56:56Z"}