Regression Tree Fields &amp;#x2014; An efficient, non-parametric approach to image labeling problems

Computer Vision – ECCV 2014

Feng

et al. 2014

Abstract. The recent advances of low-cost and mobile depth sensors dramatically extend the potential of 3D shape retrieval and analysis. While the traditional research of 3D retrieval mainly focused on searching by a rough 2D sketch or with a high-quality CAD model, we tackle a novel and challenging problem of cross-domain 3D shape retrieval, in which users can use 3D scans from low-cost depth sensors like Kinect as queries to search CAD models in the database. To cope with the imperfection of user-captured models such as model noise and occlusion, we propose a cross-domain shape retrieval framework, which minimizes the potential function of a Conditional Random Field to efficiently generate the retrieval scores. In particular, the potential function consists of two critical components: one unary potential term provides robust crossdomain partial matching and the other pairwise potential term embeds spatial structures to alleviate the instability from model noise. Both potential components are efficiently estimated using random forests with 3D local features, forming a Regression Tree Field framework. We conduct extensive experiments on two recently released user-captured 3D shape datasets and compare with several state-of-the-art approaches on the cross-domain shape retrieval task. The experimental results demonstrate that our proposed method outperforms the competing methods with a significant performance gain.

Section: R E G R E S S I O N T R E E F I E L D C O N S T R U C T I O Nmentioning

confidence: 99%

Section: Approachmentioning

confidence: 99%

Section: Formulationmentioning

confidence: 99%

See 1 more Smart Citation

From Low-Cost Depth Sensors to CAD: Cross-Domain 3D Shape Retrieval via Regression Tree Fields

Computer Vision – ECCV 2014

Feng

et al. 2014

Procedings of the British Machine Vision Conference 2013

“…Such and similar approaches have been used e.g. in computer vision for several models: fields of experts [8], regression tree fields [5] to name a few.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised learning for MRFs on bipartite graphs

Flach¹,

Sixta²

2013

We consider unsupervised (parameter) learning for general Markov random fields on bipartite graphs. This model class includes Restricted Boltzmann Machines. We show that besides the widely used stochastic gradient approximation (a.k.a. Persistent Contrastive Divergence) there is an alternative learning approach -a modified EM algorithm which is tractable because of the bipartiteness of the model graph. We compare the resulting double loop algorithm and the PCD learning experimentally and show that the former converges faster and more stable than the latter.

“…For the random-field-based priors, Jancsary et al [9] restrict the potential functions of the CRF to be Gaussian functions for faster inference. To compensate for the performance drop from limiting forms of potential functions, regression trees are trained to discriminatively determine the mean and covariance of the potential functions, resulting in a Regression Tree Field formulation [10] that provides state-of-the-art performance for denoising and inpainting. This method can be interpreted as using random forests to pre-index a flexible prior defined on cliques in the random field.…”

Section: Introductionmentioning

confidence: 99%

Discriminative Indexing for Probabilistic Image Patch Priors

Cho

Computer Vision – ECCV 2014

et al. 2014

Abstract. Newly emerged probabilistic image patch priors, such as Expected Patch Log-Likelihood (EPLL), have shown excellent performance on image restoration tasks, especially deconvolution, due to its rich expressiveness. However, its applicability is limited by the heavy computation involved in the associated optimization process. Inspired by the recent advances on using regression trees to index priors defined on a Conditional Random Field, we propose a novel discriminative indexing approach on patch-based priors to expedite the optimization process. Specifically, we propose an efficient tree indexing structure for EPLL, and overcome its training tractability challenges in high-dimensional spaces by utilizing special structures of the prior. Experimental results show that our approach accelerates state-of-the-art EPLL-based deconvolution methods by up to 40 times, with very little quality compromise.