Max Margin AND/OR Graph learning for parsing the human body

By Long Zhu, Yifei Lu, Chenxi Lin, Alan L. Yuille. 2008

95 citations recorded by the source corpus.

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

Growing Interpretable Part Graphs on ConvNets via Multi-Shot Learning (2016, methodology): “As in (Zhu et al. 2008; Si and Zhu 2013), the AOG is suitable to represent the semantic hierarchy, which enables semantic-level interactions between human and neural networks.”
Articulated Pose Estimation by a Graphical Model with Image Dependent Pairwise Relations (2014, background): “Most work on pose estimation has been based on graphical model [7, 5, 22, 21].”
Interpreting CNN knowledge via an Explanatory Graph (2017, methodology): “Inspired by [28, 25], we add an hierarchical And-Or graph to further associate implicit part patterns with explicit part names.”
Learning hierarchical poselets for human parsing (2011, methodology): “Second, the OR-nodes in [30] are conceptually similar to poselets in our case.”
Interpretable R-CNN (2017, background): “There are a variety of interpretable models based on image grammar [83, 17, 48, 82], which can offer intuitive and deep explanation, but often are suffered from difficulties in learning model structures and recently being outperformed in terms of accuracy by deep neural networks significantly.”
Articulated part-based model for joint object detection and pose estimation (2011, background): “APM is also related to grammar models for images or objects [26, 24].”
Towards Unified Human Parsing and Pose Estimation (2014, background): “The major contributions of this work include: • We build a novel Hybrid Parsing Model for unified human parsing and pose estimation.”
AOGNets: Compositional Grammatical Architectures for Deep Learning (2017, methodology): “Object detection grammar was the dominant approaches for object detection [9, 70, 52, 38, 36, 37], and has recently been integrated with DNNs [57, 58, 5].”
Latent Hierarchical Model of Temporal Structure for Complex Activity Classification (2014, methodology): “It has been widely used in computer vision and achieved successes on various tasks, such as learning feature hierarchies [38], [39], object detection [16], [40], [41], human body parsing [42], image parsing [43], and video understanding [44].”
Latent hierarchical structural learning for object detection (2010, background): “In this paper we show that simple part structures are sufficient to obtain state of the art performance.”
Towards Interpretable Object Detection by Unfolding Latent Structures (2019, methodology): “There are a variety of interpretable models based on image grammar [70, 15, 41, 69], which can offer intuitive and deep explanation, but often are suffered from difﬁculties in learning model structures and recently being outperformed in terms of accuracy by deep neural networks signiﬁcantly.”
AOGNets: Deep AND-OR Grammar Networks for Visual Recognition (2017, methodology): “In implementation, we adopt the AND-OR grammar (AOG) [33, 46, 45].”
Interactively Transferring CNN Patterns for Part Localization (2017, background): “And-Or representation: In many studies, people used AOGs to represent the hierarchical semantic hierarchy of objects or scenes [22, 11].”
From Red Wine to Red Tomato: Composition with Context (2017, background): “The parts-based systems such as Deformable Part Models [15], grammars [25, 44, 59, 62, 67, 69], and AND-OR graphs [54, 61, 66] also rely on compositionality of objects to build recognition systems.”
Compositional Convolutional Neural Networks: A Robust and Interpretable Model for Object Recognition Under Occlusion (2020, methodology): “d for object detection. Intuitively, the ornodes allow the model to learn dierent object/part congurations, while the and-node decomposes them into smaller components. Early works in this direction [7,75,36] relied on pre-dened parts and pre-dened graph structures. To learn the AOG model with less s”
Understanding videos, constructing plots learning a visually grounded storyline model from annotated videos (2009, methodology): “While there have been several heuristic inference algorithms for AND-OR graphs [ 22 ], we formulate an integer programming approach to obtain the storyline and matchings, and solve a relaxed version of the problem in the form of a linear program.”
Parsing occluded people by flexible compositions (2014, background): “Other work that exploits part sharing includes compositional models [36] and AND-OR graphs [35, 37].”
Parsing the Hand in Depth Images (2014, methodology): “In [36] the AND/OR graph is utilized to represent the body part configuration, the parameters of which are learned by the max-margin algorithm for structure learning.”