Grouplet: A structured image representation for recognizing human and object interactions

By Bangpeng Yao, Li Fei-Fei. 2010

371 citations recorded by the source corpus.

Subfields: 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 Computer Vision and Pattern Recognition

Citations in the corpus

A self-organizing neural network architecture for learning human-object interactions (2017, background): “Yao and Fei-Fei [37] proposed a set of sophisticated visual features called Grouplet which captures spatial organization of image patches encoded through SIFT descriptors [38].”
Unsupervised Discovery of Mid-Level Discriminative Patches (2012, background): “Moreover, discovering these second-order relationships can provide us with “doublets” [34] (which could be further generalized to grouplets [22,35]) that can themselves be highly discriminative and useful as mid-level features in their own right.”
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]).”
Weakly-Supervised Learning of Visual Relations (2017, background): “This task was studied for the special case of human-object interactions [8, 9, 17, 38, 39, 48, 49, 50], where the triplet is in the form ( person, action, object ) .”
Automatic Description Generation from Images: A Survey of Models, Datasets, and Evaluation Measures (2016, neutral): “…attributes (e.g., Lampert, Nickisch, & Harmeling, 2009; Berg, Berg, & Shih, 2010; Parikh & Grauman, 2011), compute scene properties (e.g., Oliva & Torralba, 2001; Lazebnik, Schmid, & Ponce, 2006), and recognize human-object interactions (e.g., Prest, Schmid, & Ferrari, 2012; Yao & Fei-Fei, 2010).”
From Lifestyle Vlogs to Everyday Interactions (2017, background): “This area has been extensively studied in vision, typically aiming to infer affordances in still images [12, 14, 32, 47, 46], understand scenes by observing them over time [9, 6], or use them as a building block for scene understanding [20].”
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].”
Mining Motion Atoms and Phrases for Complex Action Recognition (2013, methodology): “AND/OR structure have been successfully used in object and action classiﬁcation tasks [29, 3].”
Modeling mutual context of object and human pose in human-object interaction activities (2010, background): “For simplicity, we assume that only one object is involved in each activity.”
Learning Human-Object Interactions by Graph Parsing Neural Networks (2018, methodology): “Early work in HOI understanding studied Bayesian model [15, 16], utilized contextual relationship between human and objects [47–49], learned structured representations with spatial interaction and context [8], exploited compositional models [9], or referred to a set of HOI exemplars [19].”
Combining randomization and discrimination for fine-grained image categorization (2011, methodology): “In this section, we evaluate our algorithm on three finegrained image datasets: the action classification dataset in PASCAL VOC2010 [7] (Sec.5.1), actions of peopleplaying-musical-instrument (PPMI) [22] (Sec.5.2), and a subordinate object categorization dataset of 200 bird species [21] (Sec.5.3).”
HMDB: A large video database for human motion recognition (2011, background): “Clips extracted from the same movie usually have similar scenes.”
Web-Based Classifiers for Human Action Recognition (2012, background): “Recent work of Yao and Li [30] has looked into the relationship between poses and objects in still images.”
Action Recognition with Exemplar Based 2.5D Graph Matching (2012, methodology): “We carry out experiments on two publicly available datasets: the people playing musical instrument (PPMI) dataset [37] and the PASCAL VOC 2011 action classification dataset [10].”
Beyond Actions: Discriminative Models for Contextual Group Activities (2010, background): “In still image action recognition, object-action context [6, 9, 23, 24] is a popular type of context used for human-object interaction.”
Sherlock: Scalable Fact Learning in Images (2015, methodology): “We began our data collection by augmenting existing datasets with fact language view labels fl: PPMI [41], Stanford40 [42], Pascal Actions [11], Sports [16], Visual Phrases [34], INTERACT [4] datasets.”
Human-object interaction recognition by learning the distances between the object and the skeleton joints (2015, background): “In another work of Yao et al [9], it treated object and human pose as the context of each other in human-object interaction activities.”
“Important stuff, everywhere!” Activity recognition with salient proto-objects as context (2014, background): “One automatic approach to mine discriminative image regions based on their strong correlations with the target class is the grouplet [41].”
Hand detection using multiple proposals (2011, background): “Our motivation for this is that having a reliable hand detector facilitates many other tasks in human visual recognition, such as determining human layout [2, 13, 17, 24] and actions from static images [10, 13, 19, 29].”