Max-Margin Early Event Detectors

By Minh Hoai Nguyen, Fernando De la Torre. 2012

363 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language (2022, background): “In the research community, activity forecasting has been often formulated as an extension of action classification, tracking, or feature generation: Given a sequence of image frames, they directly predict a few categorized actions [124, 125, 126], human locations [127], or image features [128] to be”
Skeleton-Based Online Action Prediction Using Scale Selection Network (2019, methodology): “[20], pedestrian detection [21], activity detection [22], [23], [24], [25], etc.”
RGB-D-based Human Motion Recognition with Deep Learning: A Survey (2017, methodology): “To generalize R-CNN from 2D to 3D, Hou et al. (2017a) proposed an end-to-end 3D CNN-based approach for action detection in videos.”
PKU-MMD: A Large Scale Benchmark for Continuous Multi-Modal Human Action Understanding (2017, background): “Those methods are designed to capture local features from the sequences and then to classify them by traditional classifiers like Support Vector Machine (SVM).”
An Uncertain Future: Forecasting from Static Images Using Variational Autoencoders (2016, background): “For example, some works break video sequences into discrete actions, and attempt to detect the earliest frames in an action in order to predict the rest [13].”
Newtonian Image Understanding: Unfolding the Dynamics of Objects in Static Images (2015, background): “Action Recognition: Early prediction of activities has been discussed in [29, 27, 16, 23].”
Online Human Action Detection using Joint Classification-Regression Recurrent Neural Networks (2016, methodology): “In [5], a learning formulation based on a structural SVM is proposed to recognize partial events, enabling early detection.”
Dense Optical Flow Prediction from a Static Image (2015, background): “Others have looked at more semantic forms of prediction: that is, predict the action class of what is going to happen next [5, 14].”
Anticipating Human Activities Using Object Affordances for Reactive Robotic Response (2013, background): “There are also a few recent works which address the task of early recognition [22], [23].”
Inferring the Why in Images (2014, background): “There also has been work in forecasting activities [16, 32] and early event detection [11], but they are interested in predicting the future in videos while we wish to explain the motivations of actions.”
Rolling-Unrolling LSTMs for Action Anticipation from First-Person Video (2020, methodology): “To assess the timeliness of anticipations, we propose a novel evaluation measure inspired by the AMOC curve [48].”
Anticipating the future by watching unlabeled video (2015, methodology): “Ryoo [35] and Hoai and De la Torre [11] propose models to predict actions in early stages.”
Leaving Some Stones Unturned: Dynamic Feature Prioritization for Activity Detection in Streaming Video (2016, methodology): “In [19], a structured output approach learns to recognize partial events in untrimmed video.”
Exploring the Trade-off Between Accuracy and Observational Latency in Action Recognition (2012, neutral): “While including data from xt−10 and xt−30 makes our features not precisely a “pose”, we consider it a more intuitive term, as we do not use fine-grained sequence data, nor do we delay classification by looking further ahead in the data stream.”
RED: Reinforced Encoder-Decoder Networks for Action Anticipation (2017, background): “[5, 6] first proposed the problem of early event detection.”
Factorized Variational Autoencoders for Modeling Audience Reactions to Movies (2017, background): “Predicting future events/observations is well studied [15, 46], but previous work primarily focused on anticipating immediate events.”