Hypothesis testing framework for active object detection

Atanasov, Nikolay; Sankaran, Bharath; Ny, Jérôme Le; Koletschka, Thomas; Pappas, George J.; Daniilidis, Kostas

doi:10.1109/icra.2013.6631173

Cited by 22 publications

(13 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar to [9], we use mutual information to decide the NBV. However, we consider this 1 The model is precisely a convolutional DBM where all the connections are undirected, while ours is a convolutional DBN. problem at the precise voxel level allowing us to infer how voxels in a 3D region would contribute to the reduction of recognition uncertainty.…”

Section: Related Workmentioning

confidence: 99%

“…However this multi-view problem is intrinsically 3D in nature. Atanasov et al, [1,2] implement the idea in real world robots, but they assume that there is only one object associated with each class reducing their problem to instance-level recognition with no intra-class variance. Similar to [9], we use mutual information to decide the NBV.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

3D ShapeNets: A deep representation for volumetric shapes

Song

Khosla

et al. 2015

2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

1,722

440

View full text Add to dashboard Cite

Abstract3D shape is a crucial but heavily underutilized cue in today's computer vision systems, mostly due to the lack of a good generic shape representation. With the recent availability of inexpensive 2.5D depth sensors (e.g. Microsoft Kinect), it is becoming increasingly important to have a powerful 3D shape representation in the loop. Apart from category recognition, recovering full 3D shapes from viewbased 2.5D depth maps is also a critical part of visual understanding. To this end, we propose to represent a geometric 3D shape as a probability distribution of binary variables on a 3D voxel grid, using a Convolutional Deep Belief Network. Our model, 3D ShapeNets, learns the distribution of complex 3D shapes across different object categories and arbitrary poses from raw CAD data, and discovers hierarchical compositional part representations automatically. It naturally supports joint object recognition and shape completion from 2.5D depth maps, and it enables active object recognition through view planning. To train our 3D deep learning model, we construct ModelNet -a large-scale 3D CAD model dataset. Extensive experiments show that our 3D deep representation enables significant performance improvement over the-state-of-the-arts in a variety of tasks.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

3D ShapeNets: A deep representation for volumetric shapes

Song

Khosla

et al. 2015

2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

1,722

440

View full text Add to dashboard Cite

show abstract

“…Active information acquisition is a challenging problem with diverse applications including robotics tasks [14], [15], [16], [17]. It is a sequential decision making problem where agents are tasked with acquiring information about a certain process of interest (target).…”

Section: Introductionmentioning

confidence: 99%

Learning Q-network for Active Information Acquisition

Jeong

Schlotfeldt

Hassani

et al. 2019

2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Self Cite

View full text Add to dashboard Cite

In this paper, we propose a novel Reinforcement Learning approach for solving the Active Information Acquisition problem, which requires an agent to choose a sequence of actions in order to acquire information about a process of interest using on-board sensors. The classic challenges in the information acquisition problem are the dependence of a planning algorithm on known models and the difficulty of computing information-theoretic cost functions over arbitrary distributions. In contrast, the proposed framework of reinforcement learning does not require any knowledge on models and alleviates the problems during an extended training stage. It results in policies that are efficient to execute online and applicable for real-time control of robotic systems. Furthermore, the state-of-the-art planning methods are typically restricted to short horizons, which may become problematic with local minima. Reinforcement learning naturally handles the issue of planning horizon in information problems as it maximizes a discounted sum of rewards over a long finite or infinite time horizon. We discuss the potential benefits of the proposed framework and compare the performance of the novel algorithm to an existing information acquisition method for multi-target tracking scenarios.1 These authors are with the department

show abstract

“…For example, Hager and Wegbreit (2011) demonstrate the utility of considering a prior 3-D scene model and its potential evolution over scenes. Using this observation as a premise, active perception approaches (e.g., Eidenberger and Scharinger (2010); Velez et al (2012); Atanasov et al (2013)) seek the next best view (camera pose) where previously-occluded objects may be visible, typically by formulating the problem as a partially-observable Markov decision process. Because the focus is on planning instead of estimation, this line of work is complementary to the world modeling problem, which considers estimation using measurements from an uncontrolled, arbitrary collection of camera poses.…”

Section: Semantic World Modelingmentioning

confidence: 99%

Data association for semantic world modeling from partial views

Wong

Kaelbling

Tom

et al. 2015

The International Journal of Robotics Research

View full text Add to dashboard Cite

Autonomous mobile-manipulation robots need to sense and interact with objects to accomplish high-level tasks such as preparing meals and searching for objects. To achieve such tasks, robots need semantic world models, defined as object-based representations of the world involving task-level attributes. In this work, we address the problem of estimating world models from semantic perception modules that provide noisy observations of attributes. Because attribute detections are sparse, ambiguous, and are aggregated across different viewpoints, it is unclear which attribute measurements are produced by the same object, so data association issues are prevalent. We present novel clustering-based approaches to this problem, which are more efficient and require less severe approximations compared to existing tracking-based approaches. These approaches are applied to data containing object type-and-pose detections from multiple viewpoints, and demonstrate comparable quality using a fraction of the computation time.

show abstract

Hypothesis testing framework for active object detection

Cited by 22 publications

References 39 publications

3D ShapeNets: A deep representation for volumetric shapes

3D ShapeNets: A deep representation for volumetric shapes

Learning Q-network for Active Information Acquisition

Data association for semantic world modeling from partial views

Contact Info

Product

Resources

About