Abstract: In unsupervised graph anomaly detection, existing methods usually focus on detecting outliers by learning local context information of nodes, while often ignoring the importance of global ...
Abstract: Detecting small objects in drone imagery is challenging due to low resolution and background blending, leading to limited feature information. Multiscale feature fusion can enhance detection ...