发帖
雷达卡
Neural Transformation Learning for Deep Anomaly Detection Beyond Images
Chen Qiu 1 2 Timo Pfrommer 1 Marius Kloft 2 Stephan Mandt 3 Maja Rudolph 1
Abstract formations are useful, and it is hard to design these trans-
Data transformations (e.g. rotations, reections, formations manually. This paper studies self-supervised
and cropping) play an important role in self- anomaly detection for data types beyond images. We de-
supervised lea ...
京公网安备 11010802022788号
