arXiv subjects

Huisheng Wang

Publications and source records attributed to Huisheng Wang.

23 records · Page 2Linked to original sources

arXiv2022

MovieCLIP: Visual Scene Recognition in Movies

Longform media such as movies have complex narrative structures, with events spanning a rich variety of ambient visual scenes. Domain specific challenges associated with visual scenes in movies include transitions, person coverage, and a wide array of real-life and fictional scenarios. Existing visual scene datasets in movies have limited taxonomies and don't consider the visual scene transition within movie clips. In this work, we address the problem of visual scene recognition in movies by first automatically curating a new and extensive movie-centric taxonomy of 179 scene labels derived from movie scripts and auxiliary web-based video datasets. Instead of manual annotations which can be expensive, we use CLIP to weakly label 1.12 million shots from 32K movie clips based on our proposed taxonomy. We provide baseline visual models trained on the weakly labeled dataset called MovieCLIP and evaluate them on an independent dataset verified by human raters. We show that leveraging features from models pretrained on MovieCLIP benefits downstream tasks such as multi-label scene and genre classification of web videos and movie trailers.

cs.CV↗

arXiv2021

Spatiotemporal Contrastive Video Representation Learning

We present a self-supervised Contrastive Video Representation Learning (CVRL) method to learn spatiotemporal visual representations from unlabeled videos. Our representations are learned using a contrastive loss, where two augmented clips from the same short video are pulled together in the embedding space, while clips from different videos are pushed away. We study what makes for good data augmentations for video self-supervised learning and find that both spatial and temporal information are crucial. We carefully design data augmentations involving spatial and temporal cues. Concretely, we propose a temporally consistent spatial augmentation method to impose strong spatial augmentations on each frame of the video while maintaining the temporal consistency across frames. We also propose a sampling-based temporal augmentation method to avoid overly enforcing invariance on clips that are distant in time. On Kinetics-600, a linear classifier trained on the representations learned by CVRL achieves 70.4% top-1 accuracy with a 3D-ResNet-50 (R3D-50) backbone, outperforming ImageNet supervised pre-training by 15.7% and SimCLR unsupervised pre-training by 18.8% using the same inflated R3D-50. The performance of CVRL can be further improved to 72.9% with a larger R3D-152 (2x filters) backbone, significantly closing the gap between unsupervised and supervised video representation learning. Our code and models will be available at https://github.com/tensorflow/models/tree/master/official/.

cs.CV↗

arXiv2020

Research on Intelligent Traffic Control Methods at Intersections Based on Game Theory

Based on game theory and dynamic Level-k model, this paper establishes an intelligent traffic control method for intersections, studies the influence of multi-agent vehicle joint decision-making and group behavior disturbance on system state. The simulation results show that this method has a good performance when there are more vehicles or emergency vehicles have higher priority.

eess.SY↗

arXiv2006

Branching Ratio and CP Asymmetry of $B \to ρη^{(\prime)}$ Decays in the Perturbative QCD Approach

In this paper,we calculate the branching ratios and CP-violating asymmetries for $B^0 \to ρ^0 η^{(\prime)}$ and $B^+\to ρ^+ \etap$ decays in the perturbative QCD factorization approach. In this approach, we not only calculate the usual factorizable contributions, but also evaluate the non-factorizable and annihilation type contributions. Besides the current-current operators, the contributions from the QCD and electroweak penguin operators are also taken into account. The theoretical predictions for the branching ratios are $Br(B^+ \to ρ^+ η^{(\prime)}) \approx 9 \times 10^{-6}$ and $Br(B^0 \to ρ^0 η^{(\prime)}) \approx 5 \times 10^{-8}$, which agree well with the measured values and currently available experimental upper limits. We also predict large CP-violating asymmetries in these decays: $A_{CP}^{dir}(ρ^\pm η)\approx -13 %$, $A_{CP}^{dir}(ρ^\pm η^\prime)\approx -18 %$, $A_{CP}^{dir}(ρ^0 η)\approx -41%$, $A_{CP}^{dir}(ρ^0 η^\prime)\approx -27%$, $A_{CP}^{mix}(ρ^0 η)\approx +25%$, and $A_{CP}^{mix}(ρ^0 η^\prime) \approx +11%$, which can be tested by the current or future B factory experiments.

hep-ph↗

arXiv2005

Branching Ratio and CP Asymmetry of $B \to πη^{(\prime)}$ Decays in the Perturbative QCD Approach

We calculate the branching ratios and CP-violating asymmetries for $B^0 \to π^0 η^{(\prime)}$ and $B^+\to π^+ η^{(\prime)}$ decays in the perturbative QCD (pQCD) factorization approach here. We not only calculate the usual factorizable contributions, but also evaluate the non-factorizable and annihilation type contributions. Besides the current-current operators, the contributions from the QCD and electroweak penguin operators are also taken into account. The pQCD results for the CP-averaged branching ratios are $Br(B^+ \to π^+ η) \approx 4.1 \times 10^{-6}$, $Br(B^+ \to π^+ η^\prime) \approx 2.4 \times 10^{-6}$, and $Br(B^0 \to π^0 η^{(\prime)}) \approx 0.2 \times 10^{-6}$, which agree very well with the measured values or currently available experimental upper limits. We also predict large CP-violating asymmetries in these decays: $A_{CP}^{dir}(π^\pm \etap)\sim A_{CP}^{dir}(π^0 \etap)\sim -0.35 $, and $A_{CP}^{mix}(π^0 \etap)\sim 0.67$, but with large errors. The pQCD prediction for $A_{CP}^{dir}(π^\pm η)$ ($A_{CP}^{dir}(π^\pm η^\prime))$ has the same (opposite) sign with the primary measured values. Further improvements in both theory and experiments are needed to clarify this discrepancy.

hep-ph↗