We consider two less-emphasized temporal properties of video: 1. Temporal cues are fine-grained; 2. Temporal modeling needs reasoning. To tackle both problems at once, we exploit approximated bilinear modules (ABMs) for temporal modeling. There are two main points making the modules effective: two-layer MLPs can be seen as a constraint approximation of bilinear operations, thus can be used to construct deep ABMs in existing CNNs while reusing pretrained parameters; frame features can be divided into static and dynamic parts because of visual repetition in adjacent frames, which enables temporal modeling to be more efficient. Multiple ABM variants and implementations are investigated, from high performance to high efficiency. Specifically, we show how two-layer subnets in CNNs can be converted to temporal bilinear modules by adding an auxiliary-branch. Besides, we introduce snippet sampling and shifting inference to boost sparse-frame video classification performance. Extensive ablation studies are conducted to show the effectiveness of proposed techniques. Our models can outperform most state-of-the-art methods on Something-Something v1 and v2 datasets without Kinetics pretraining, and are also competitive on other YouTubelike action recognition datasets. Our code is available on https://github.com/zhuxinqimac/abm-pytorch.
Metamaterials are composed of periodic or quasi‐periodic subwavelength structures, having electric and/or magnetic responses. Metamaterials can arbitrarily tailor the refractive index by artificially tailoring the unit‐cell geometries and dimensions. Recent years, as the two dimensional equivalent of bulk metamaterials, metasurfaces have caused considerable attentions due to the lower profile and simpler to fabricate than bulk metamaterials. Metasurfaces can impart discontinuities on electromagnetic wavefronts and can achieve the arbitrary transmission phase of the whole period range. Metamaterials and metasurfaces have led to the realizations of novel electromagnetic properties and functionalities through tailoring subwavelength structures and integrating functional materials. In this letter, two planar lenses are respectively proposed to control the beam direction of the horn antenna by using a gradient refractive index (GRIN) metamaterial and a gradient phase (GRPH) metasurface. It is shown that the antenna beam direction can be steered by controlling the refractive index of the GRIN metamaterial or the transmission phase of the GRPH metasurface. The differences between the two planar lenses for controlling the antenna beam are illustrated.
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.