Ieee Transactions on Pattern Analysis an Machine Intelligence Structured Learning of Human Interactions in Tv Shows

By Alonso Patron-Perez, Marcin Marszalek, Ian David Reid, Andrew Zisserman

How researchers see this paper

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

Citations in the corpus

Dyadic Interaction Detection from Pose and Flow (2014, methodology): “[18], who focused on dyadic interactions such as hand shakes and hugs.”
A Review of Human Activity Recognition Methods (2015, methodology): “…focus has also been given to recognizing human activities from real life videos, such as movies and TV shows, by exploiting scene contexts to localize activities and understand human interactions (Marszałek et al., 2009; Patron-Perez et al., 2012; Bojanowski et al., 2013; Hoai and Zisserman, 2014).”
Detecting People Looking at Each Other in Videos (2013, unknown): “shaking hands) (Patron-Perez et al. 2012; Yang et al. 2012).”
A Hierarchical Representation for Future Action Prediction (2014, methodology): “We deﬁne the starting point of an action according to the annotation of the dataset [14].”
A survey of video datasets for human action and activity recognition (2013, methodology): “between two people [166], activity recognition [41,115]”
Context-Aware CNNs for Person Head Detection (2015, methodology): “The combination of our context-aware models and the CNN-based local detector achieves state-of-the-art results on our dataset and the two existing human detection datasets, TVHI and Casablanca.”
Aligning plot synopses to videos for story-based retrieval (2015, background): “The recent advances in this domain are mainly in popular computer vision topics such as action recognition [21], person identification [6,10,37], pose estimation [5], and human interaction analysis [30].”
Multi-task Recurrent Neural Network for Immediacy Prediction (2015, methodology): “Previous related benchmark datasets: the WeAreFamily dataset [7] is used for multiple people pose estimation; The TVHI dataset (video) [23] and TVHI+ dataset (video) [13] are used for interaction classification.”
A Hybrid Loss for Multiclass and Structured Prediction (2014, background): “Our feature representation is a combination of several visual cues including multiclass SVM action classiﬁcation scores, human body poses and the relative position between two individuals, which have been exploited to distinguish different actions in [14, 13].”
Attention Transfer from Web Images for Video Recognition (2017, methodology): “We use the 200 positive videos, and keep the train/test split as [28].”