Recognizing proxemics in personal photos

By Yi Yang, Simon Baker, Anitha Kannan, Deva Ramanan. 2012

95 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

Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations (2016, background): “Relationships have already shown their utility in improving visual cognitive tasks (Antol et al. 2014; Yang et al. 2012).”
DeepID-Net: multi-stage and deformable deep convolutional neural networks for object detection (2014, background): “The context information inves tigated in previous works includes regions surrounding objects [9, 12, 22], object-scene interaction [ 13], and the presence, location, orientation and size relationship among ob jects [3, 57, 58, 11, 41, 22, 49, 13, 61, 12, 59, 40, 10, 45, 51].”
A Domain Based Approach to Social Relation Recognition (2017, background): “A relevant topic is intimacy prediction based on interactive poses [10, 11].”
Zero-Shot Learning via Visual Abstraction (2014, methodology): “While some datasets exist for a small number of these categories [26], we collected a new dataset with a significantly larger number of visual classes, INTERACT.”
Human Pose Regression by Combining Indirect Part Detection and Contextual Information (2017, unknown): “To handle dif-ﬁcult cases of occlusion or partial visualization, contextual information is usually needed to provide visual cues that can be extracted from a broad region around the part location [15] or by interaction among detected parts [45].”
DeepID-Net: Object Detection with Deformable Part Based Convolutional Neural Networks (2017, unknown): “object-scene interaction [60], [61], and the presence, location, orientation and size relationship among objects [33], [58], [59], [60], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72].”
3D Human pose estimation: A review of the literature and analysis of covariates (2016, background): “image [15] and at the same time improves the pose estimation accuracy since it addresses occlusions between body parts; (viii) Estimating the anthropometry of a human from a single image [16, 17, 18]; (ix) 3D Avatar creation [19, 20] or controlling a 3D Avatar in games [21]; (x) Understanding the ca”
Single-Pedestrian Detection Aided by Multi-pedestrian Detection (2013, background): “The context investigated in previous works includes regions surrounding objects [4, 6, 15], object-scene interaction [7], and the presence, location, orientation and size relationship among objects [2, 38, 39, 5, 25, 15, 29, 7, 42, 6, 40, 24].”
Panoptic Studio: A Massively Multiview System for Social Interaction Capture (2016, background): “In [46], due to the relatively unreliable 2D pose detection cue [47], the motion cues from trajectory stream play a core role to reconstruct valid parts.”
Interactive body part contrast mining for human interaction recognition (2014, methodology): “Inspired by [13] using a proxemic-specific model to represent the interactions on photos, we propose an interactive body part model to represent interaction sequences.”
Computer Vision for Autonomous Vehicles: Problems, Datasets and State-of-the-Art (2017, background): “Usually, a separate context model is learned to handle occlusions, see Hoiem et al. (2008); Tu & Bai (2010); Desai et al. (2011); Yang et al. (2012).”
DeepID-Net: Deformable deep convolutional neural networks for object detection (2014, background): “The context information investigated in literature includes regions surrounding objects [5, 8, 13], object-scene interaction [9, 20], and the presence, location, orientation and size relationship among objects [2, 48, 49, 7, 31, 13, 40, 9, 53, 8, 50, 30, 6, 35, 45].”
Robust Face Landmark Estimation under Occlusion (2013, background): “These methods are effective at detecting articulated objects [44] and localizing objects from multiple views in difficult scenarios [41, 47].”
Detecting People Looking at Each Other in Videos (2013, background): “Determining whether characters are interacting using their eyelines is another step towards a fuller video understanding, and complements recent work on automatic character identification [7, 11, 37], human pose estimation [1, 6, 17, 33, 46], human action recognition [21, 22, 24, 30, 32], and social”
Multi-context Attention for Human Pose Estimation (2017, background): “Multiple Contextual Information The contextual information is generally referred to as regions surrounding the target locations [11, 13, 33], object-scene relationships [20, 18, 12], and object-object interactions [43].”
Panoptic Studio: A Massively Multiview System for Social Motion Capture (2015, background): “Since then, photographs have been–and continue to be–a fundamental tool in studying social behavior [19, 21, 13, 41, 34, 9, 1].”
AutoCaption: Automatic caption generation for personal photos (2014, background): “Currently our system runs face detection [26, 20], face recognition [4], proxemics detection [24], scene classification [23] and landmarks recognition [2].”
Multi-task Recurrent Neural Network for Immediacy Prediction (2015, methodology): “Interaction pattern is called proxemics and was estimated using deformable part based models [13] and flexible mixtures of part models in [37].”