Action recognition from a distributed representation of pose and appearance

By Subhransu Maji, Lubomir D. Bourdev, Jitendra Malik. 2011

338 citations recorded by the source corpus.

Subfields: Biomedical Engineering, 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 Biomedical Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Cross-View Action Modeling, Learning, and Recognition (2014, methodology): “[14] uses the activations of poselets, which is is a set of pose detectors.”
Contextual Action Recognition with R*CNN (2015, methodology): “[20] train action specific poselets and for each instance create a poselet activation vector that is classified using SVMs.”
Visual Relationship Detection with Language Priors (2016, background): “There have been numerous eﬀorts in human-object interaction [20,21,22] and action recognition [23] to learn discriminative models that distinguish be-tween relationships where object 1 is a human ( e.g. “playing violin” [24]).”
Anticipating Human Activities Using Object Affordances for Reactive Robotic Response (2013, background): “In recent years, much effort has been made to detect human activities from still images as well as videos (e.g., [11], [12]).”
A Hierarchical Pose-Based Approach to Complex Action Understanding Using Dictionaries of Actionlets and Motion Poselets (2016, background): “Their approach has been applied to action recognition in still images [19], as well as in videos [26, 33, 41].”
Recognizing Human-Object Interactions in Still Images by Modeling the Mutual Context of Objects and Human Poses (2012, neutral): “Recognizing human activities in still images is a new problem, yet it has received much attention in recent years [14], [15], [42], [43], [44], [45], [46].”
Using k-Poselets for Detecting People and Localizing Their Keypoints (2014, background): “A poselet’s training set is associated with a consistent keypoint configuration (unlike in DPM training), so by design when there is a strong activation of a poselet in a test image, one can make considerably tighter predictions that are useful for analyzing attributes [4], segmenting objects [6], r”
R-CNNs for Pose Estimation and Action Detection (2014, methodology): “[19] train action specific poselets and for each instance create a poselet activation vector that is being classified using SVMs.”
Multi-Modal Knowledge Graph Construction and Application: A Survey (2022, methodology): “In early studies, action is defined as a serial of poses of different parts of the body, and the discriminative features are mined by heuristics approaches [73], [74], [75].”
iCAN: Instance-Centric Attention Network for Human-Object Interaction Detection (2018, background): “Over the past few years, there has been rapid progress in visual recognition tasks, including object detection [8, 10, 27, 34], segmentation [5, 11, 18, 28], and action recognition [6, 9, 13, 30, 41].”
CIDEr: Consensus-based image description evaluation (2014, background): “Recent advances in object recognition [15], attribute classification [23], action classification [26, 9] and crowdsourcing [40] have increased the interest in solving higher level scene understanding problems.”
Actions and Attributes from Wholes and Parts (2014, methodology): “Action Poselets [20] is a part based approach using HOG features, while Oquab et al.”
Recognize complex events from static images by fusing deep channels (2015, background): “Analyzing human actions [10] with the help of human poses [21, 38] and human-object interactions [2, 40] also provides significant cues in recognizing certain categories of events.”
Attentional Pooling for Action Recognition (2017, background): “Traditional approaches to action recognition relied on object detection [11, 19, 57], articulated pose [29, 34, 35, 55, 57], dense trajectories [52, 53] and part-based/structured models [9, 56, 58].”
A survey on still image based human action recognition (2014, methodology): “They sampled 100 windows from each image-based on the candidates from objectness measure [56] where feature vectors can be extracted from each of the 100 windows.”
Exploring discriminative pose sub-patterns for effective action classification (2013, background): “Another recent research trend to employ pose as a cue in action recognition is closely related to the concept of “poselets” [22, 12, 32].”
Human Action Recognition With Video Data: Research and Evaluation Challenges (2014, background): “books, or dictionary-based approaches [50], [60], [71], [73], [75], [76], [106]–[114] apply the same principle since they consider the image features from the complete video segments.”