Contemporary deep learning multi-scale deblurring models suffer from many issues: 1) They perform poorly on non-uniformly blurred images/videos; 2) Simply increasing the model depth with finer-scale levels cannot improve deblurring; 3) Individual RGB frames contain a limited motion information for deblurring; 4) Previous models have a limited robustness to spatial transformations and noise.
The goal of multi-level feature design is to extract feature representations at different layer-wise levels of CNN, realizing several levels of visual abstraction to achieve robust few-shot learning.
Our layer combines the feature vectors and their respective spatial locations in the feature maps produced by the last convolutional layer of CNN into a positive definite matrix with second-order statistics to which PN operators are applied, forming so-called Second-order Pooling (SOP).
The heavy defocus blur in DP pairs affects the performance of matching-based depth estimation approaches.
Such encoded blocks are aggregated by permutation-invariant pooling to make our approach robust to varying action lengths and long-range temporal dependencies whose patterns are unlikely to repeat even in clips of the same class.
Ranked #6 on Few Shot Action Recognition on Kinetics-100
The majority of existing few-shot learning methods describe image relations with binary labels.
In this paper, we study the impact of scale and location mismatch in the few-shot learning scenario, and propose a novel Spatially-aware Matching (SM) scheme to effectively perform matching across multiple scales and locations, and learn image relations by giving the highest weights to the best matching pairs.
To the best of our knowledge, we are the first to leverage saliency maps for such a task and we demonstrate their usefulness in hallucinating additional datapoints for few-shot learning.
depth, we propose a stacked version of our multi-patch model.
Ranked #7 on Deblurring on RealBlur-R (trained on GoPro) (SSIM (sRGB) metric)
In this paper, we propose a similarity learning network leveraging second-order information and Power Normalizations.
Specifically, we leverage two sources of datapoints (observed and auxiliary) to train some classifier to recognize which test datapoints come from seen and which from unseen classes.
To achieve robust baselines, we build on a recent approach that aligns per-class scatter matrices of the source and target CNN streams.
In this paper, we reconsider these operators in the deep learning setup by introducing a novel layer that implements PN for non-linear pooling of feature maps.
To achieve robust baselines, we build on a recent approach that aligns per-class scatter matrices of the source and target CNN streams .
no code implementations • 25 Jul 2017 • Changbo Fu, Xiaopeng Zhou, Xun Chen, Yunhua Chen, Xiangyi Cui, Deqing Fang, Karl Giboni, Franco Giuliani, Ke Han, Xingtao Huang, Xiangdong Ji, Yonglin Ju, Siao Lei, Shaoli Li, Huaxuan Liu, Jianglai Liu, Yugang Ma, Yajun Mao, Xiangxiang Ren, Andi Tan, Hongwei Wang, Jimin Wang, Meng Wang, Qiuhong Wang, Siguang Wang, Xuming Wang, Zhou Wang, Shiyong Wu, Mengjiao Xiao, Pengwei Xie, Binbin Yan, Yong Yang, Jianfeng Yue, Hongguang Zhang, Tao Zhang, Li Zhao, Ning Zhou
We report new searches for the solar axions and galactic axion-like dark matter particles, using the first low-background data from PandaX-II experiment at China Jinping Underground Laboratory, corresponding to a total exposure of about $2. 7\times 10^4$ kg$\cdot$day.
High Energy Physics - Experiment Solar and Stellar Astrophysics High Energy Physics - Phenomenology