Recognizing 50 human action categories of web videos

By Kishore K. Reddy, Mubarak Shah. 2013

694 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Human-Computer Interaction

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Human-Computer Interaction

Citations in the corpus

YouTube2Text: Recognizing and Describing Arbitrary Activities Using Semantic Hierarchies and Zero-Shot Recognition (2013, methodology): “Introduction Despite recent advances in activity recognition [27, 24, 2, 5] automatic understanding and description of activities “in-the-wild” remains challenging.”
Projective dictionary pair learning for pattern classification (2014, methodology): “We evaluate the proposed DPL method on various visual classification datasets, including two face databases (Extended YaleB [22] and AR [24]), one object categorization database (Caltech101) [25], and one action recognition database (UCF 50 action [26]).”
Simple to Complex Transfer Learning for Action Recognition (2016, methodology): “Following these rules, we design a relationship matrix F on the Olympic Sports dataset [25] and UCF50 dataset [26].”
Generating Natural-Language Video Descriptions Using Text-Mined Knowledge (2013, methodology): “During this filtering, we also allow synonyms of these object names by including all words with a Lesk similarity (as implemented by Pedersen et al. (2004)) of at least 0.5.1 Using this approach, we chose 235 potential test videos; the remaining 1,735 videos were reserved for training.”
UTD-MHAD: A multimodal dataset for human action recognition utilizing a depth camera and a wearable inertial sensor (2015, background): “There are RGB camera-based datasets such as KTH [1] and UCF50 [3] that have facilitated the comparison of different recognition approaches.”
Human Action Recognition across Datasets by Foreground-Weighted Histogram Decomposition (2014, methodology): “The gist descriptor was computed for every 50th frame of all videos in the KTH [20], UCF50 [16], and HMDB51 [11] datasets.”
The THUMOS challenge on action recognition for videos "in the wild" (2016, neutral): “The next in the series were UCF11 ( Liu et al., 2009 ) and UCF50 ( Reddy and Shah, 2013 ), released in 2009 and 2011, respectively.”
Efﬁcient feature extraction, encoding and classiﬁcation for action recognition (positive): “We evaluate the speed and accuracy of our approach on Hollywood-2 [19], UCF50 [27], HMDB51 [14] and UT-Interaction [30] benchmarks.”
Multi-Modal Domain Adaptation for Fine-Grained Action Recognition (2020, methodology): “The approaches [7, 19, 36] are evaluated on 5-7 pairs of domains from subsets of coarse-grained action recognition and gesture datasets, for example aligning UCF [41] to Olympics [34].”
Mining Motion Atoms and Phrases for Complex Action Recognition (2013, methodology): “We conduct experiments on two complex action datasets: Olympic Sports dataset [15] and UCF50 dataset [18].”
Automatic Construction of Action Datasets Using Web Videos with Density-Based Cluster Analysis and Outlier Detection (2015, methodology): “This result is comparable to other approaches on the same dataset [27, 18].”