Collaborative Discrepancy Optimization for Reliable Image Anomaly Localization
نویسندگان
چکیده
Most unsupervised image anomaly localization methods suffer from overgeneralization because of the high generalization abilities convolutional neural networks, leading to unreliable predictions. To mitigate overgeneralization, this study proposes collaboratively optimize normal and abnormal feature distributions with assistance synthetic anomalies, namely collaborative discrepancy optimization (CDO). CDO introduces a margin module an overlap two key factors determining performance, i.e. , between (DDs) samples. With CDO, large small DDs are obtained, prediction reliability is boosted. Experiments on MVTec2D MVTec3D show that effectively mitigates achieves great performance real-time computation efficiency. A real-world automotive plastic parts inspection application further demonstrates capability proposed CDO.
منابع مشابه
Post: a Decentralized Platform for Reliable Collaborative Applications Post: a Decentralized Platform for Reliable Collaborative Applications
POST: A Decentralized Platform for Reliable Collaborative Applications
متن کاملCollaborative Alerts Ranking for Anomaly Detection
Given a large number of low-level heterogeneous categorical alerts from an anomaly detection system, how to characterize complex relationships between different alerts, filter out false positives, and deliver trustworthy rankings and suggestions to end users? This problem is motivated by and generalized from applications in enterprise security and attack scenario reconstruction. While existing ...
متن کاملDiscrepancy-based Evolutionary Diversity Optimization
Diversity plays a crucial role in evolutionary computation. While diversity has been mainly used to prevent the population of an evolutionary algorithm from premature convergence, the use of evolutionary algorithms to obtain a diverse set of solutions has gained increasing attention in recent years. Diversity optimization in terms of features on the underlying problem allows to obtain a better ...
متن کاملSemidefinite optimization in discrepancy theory
Recently, there have been several newdevelopments in discrepancy theory based on connections to semidefinite programming. This connection has been useful in several ways. It gives efficient polynomial time algorithms for several problems for which only non-constructive results were previously known. It also leads to several new structural results in discrepancy itself, such as tightness of the ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: IEEE Transactions on Industrial Informatics
سال: 2023
ISSN: ['1551-3203', '1941-0050']
DOI: https://doi.org/10.1109/tii.2023.3241579