深度学习之异常检测+视频预测+预训练+视频标注+镜头边界检测+行为分割+视频生成gif--附带源码和作者主页

异常检测

 

Fully Convolutional Neural Network for Fast Anomaly Detection in Crowded Scenes

• arxiv: http://arxiv.org/abs/1609.00866

Anomaly Detection in Video Using Predictive Convolutional Long Short-Term Memory Networks

• intro: Rochester Institute of Technology
• arxiv: https://arxiv.org/abs/1612.00390

Abnormal Event Detection in Videos using Spatiotemporal Autoencoder

• arxiv: https://arxiv.org/abs/1701.01546
• github: https://github.com/yshean/abnormal-spatiotemporal-ae

Abnormal Event Detection in Videos using Generative Adversarial Nets

• intro: Best Paper / Student Paper Award Finalist, IEEE International Conference on Image Processing (ICIP), 2017
• arxiv: https://arxiv.org/abs/1708.09644

Joint Detection and Recounting of Abnormal Events by Learning Deep Generic Knowledge

• intro: ICCV 2017
• arxiv: https://arxiv.org/abs/1709.09121

An overview of deep learning based methods for unsupervised and semi-supervised anomaly detection in videos

• intro: Uncanny Vision Solutions
• arxiv: https://arxiv.org/abs/1801.03149

STAN: Spatio-Temporal Adversarial Networks for Abnormal Event Detection

• intro: ICASSP 2018
• arxiv: https://arxiv.org/abs/1804.08381

Video Anomaly Detection and Localization via Gaussian Mixture Fully Convolutional Variational Autoencoder

• https://arxiv.org/abs/1805.11223

 

 

 

视频预测

 

Deep multi-scale video prediction beyond mean square error

• intro: ICLR 2016
• arxiv: http://arxiv.org/abs/1511.05440
• github: https://github.com/coupriec/VideoPredictionICLR2016
• github(TensorFlow): https://github.com/dyelax/Adversarial_Video_Generation
• demo: http://cs.nyu.edu/~mathieu/iclr2016.html

Unsupervised Learning for Physical Interaction through Video Prediction

• intro: NIPS 2016
• arxiv: https://arxiv.org/abs/1605.07157
• github: https://github.com/tensorflow/models/tree/master/video_prediction

Generating Videos with Scene Dynamics

• intro: NIPS 2016
• intro: The model learns to generate tiny videos using adversarial networks
• project page: http://web.mit.edu/vondrick/tinyvideo/
• paper: http://web.mit.edu/vondrick/tinyvideo/paper.pdf
• github: https://github.com/cvondrick/videogan

 

 

 

预训练

 

Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning

• project page: https://coxlab.github.io/prednet/
• arxiv: http://arxiv.org/abs/1605.08104
• github: https://github.com/coxlab/prednet
• github: https://github.com/e-lab/torch-prednet

Diversity encouraged learning of unsupervised LSTM ensemble for neural activity video prediction

• arxiv: https://arxiv.org/abs/1611.04899

Video Ladder Networks

• inro: NIPS 2016 workshop on ML for Spatiotemporal Forecasting
• arxiv: https://arxiv.org/abs/1612.01756

Unsupervised Learning of Long-Term Motion Dynamics for Videos

• intro: Stanford University
• arxiv: https://arxiv.org/abs/1701.01821

One-Step Time-Dependent Future Video Frame Prediction with a Convolutional Encoder-Decoder Neural Network

• intro: NCCV 2016
• arxiv: https://arxiv.org/abs/1702.04125

Fully Context-Aware Video Prediction

• intro: ETH Zurich & NNAISENSE
• keywords: unsupervised learning through video prediction, Parallel Multi-Dimensional LSTM
• project page: https://sites.google.com/view/contextvp
• arxiv: https://arxiv.org/abs/1710.08518

Novel Video Prediction for Large-scale Scene using Optical Flow

• intro: University of Victoria & Tongji University & Horizon Robotics
• arxiv: https://arxiv.org/abs/1805.12243

 

 

 

视频标注

 

Automatic Image and Video Tagging

• blog: http://scottge.net/2015/06/30/automatic-image-and-video-tagging/

Tagging YouTube music videos with deep learning - Alexandre Passant

• keywords: Clarifai's deep learning API
• blog: http://apassant.net/2015/07/03/tagging-youtube-music-clarifai-deep-learning/

 

 

 

镜头边界检测

Large-scale, Fast and Accurate Shot Boundary Detection through Spatio-temporal Convolutional Neural Networks

https://arxiv.org/abs/1705.03281

Ridiculously Fast Shot Boundary Detection with Fully Convolutional Neural Networks

• intro: obtains state-of-the-art results while running at an unprecedented speed of more than 120x real-time.
• arxiv: https://arxiv.org/abs/1705.08214

 

 

 

视频行为分割

TricorNet: A Hybrid Temporal Convolutional and Recurrent Network for Video Action Segmentation

• intro: University of Rochester
• arxiv: https://arxiv.org/abs/1705.07818

 

 

 

视频生成gif

Video2GIF: Automatic Generation of Animated GIFs from Video (Robust Deep RankNet)

• intro: 3D CNN, ranking model, Huber loss, 100K GIFs/video sources dataset
• arxiv: http://arxiv.org/abs/1605.04850
• github(dataset): https://github.com/gyglim/video2gif_dataset
• results: http://video2gif.info/
• demo site: http://people.ee.ethz.ch/~gyglim/work_public/autogif/
• review: http://motherboard.vice.com/read/these-fire-gifs-were-made-by-artificial-intelligence-yahoo

Creating Animated GIFs Automatically from Video

https://yahooresearch.tumblr.com/post/148009705216/creating-animated-gifs-automatically-from-video