We tackle the task of semi-supervised video object segmentation, i.e, pixel-level object classification of the images in video sequences using very limited ground truth training data of its corresponding video. Recently introduced online adaptation of convolutional neural networks for video object segmentation (OnAVOS) has achieved good results by pretraining the network, fine-tuning on the first frame and training the network at test time using its approximate prediction as newly obtained ground truth. We propose Flow Adaptive Video Object Segmentation (FAVOS) that refines the generated adaptive ground truth for online updates and utilizes temporal consistency between video frames with the help of optical flow. We validate our approach on the DAVIS Challenge and achieve rank 1 results on the DAVIS 2016 Challenge (single-object segmentation) and competitive scores on both DAVIS 2018 Semi-supervised Challenge and Interactive Challenge (multi-object segmentation). While most models tend to have increasing complexity for the challenging task of video object segmentation, FAVOS provides a simple and efficient pipeline that produces accurate predictions.
Nowadays the yttrium iron garnet (Y3Fe5O12, YIG) films are widely used in the microwave and spin wave devices due to their low damping constant and long propagation distance for spin waves. However, the performances, especially the frequency stability, are seriously affected by the relaxation of the interface magnetic moments. In this study, the effect of out-of-plane magnetization depinning on the resonance frequency shift (Δf
r) was investigated for 3-μm YIG films grown on Gd3Ga5O12 (GGG) (111) substrates by liquid-phase epitaxy. It is revealed that the ferromagnetic resonance (FMR) and spin wave propagation exhibit a very slow relaxation with relaxation time τ even longer than one hour under an out-of-plane external magnetic bias field. The Δf
r span of 15.15–24.70 MHz is observed in out-of-plane FMR and forward volume spin waves. Moreover, the Δf
r and τ depend on the magnetic field. The Δf
r can be attributed to that the magnetic moments break away from the pinning layer at the YIG/GGG interface. The thickness of the pinning layer is estimated to be about 9.48 nm to 15.46 nm according to the frequency shifting. These results indicate that Δf
r caused by the pinning layer should be addressed in the design of microwave and spin wave devices, especially in the transverse magnetic components.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.