Interpreting Terrestrial Images of Urban Scenes Using Discriminative Random Fields
نویسنده
چکیده
We investigate Discriminative Random Fields (DRF) which provide a principled approach for combining local discriminative classifiers that allow the use of arbitrary overlapping features, with adaptive data-dependent smoothing over the label field. We discuss the differences between a traditional Markov Random Field (MRF) formulation and the DRF model, and compare the performance of the two models and an independent sitewise classifier. Further, we present results suggesting the potential for performance enhancement by improving state of the art parameter learning methods. Eventually, we demonstrate the application feasibility on both synthetic and natural images.
منابع مشابه
Top down saliency estimation via superpixel-based discriminative dictionaries
Predicting where humans look in images has gained significant popularity in recent years. In this work, we present a novel method for learning top-down visual saliency, which is well-suited to locate objects of interest in complex scenes. During training, we jointly learn a superpixel based class-specific dictionary and a Conditional Random Field (CRF). While using such a discriminative diction...
متن کاملHierarchical and spatial structures for interpreting images of man made scenes using graphical models
Hierarchical and Spatial Structures for Interpreting Images of Man-made Scenes Using Graphical Models The task of semantic scene interpretation is to label the regions of an image and their relations into meaningful classes. Such task is a key ingredient to many computer vision applications, including object recognition, 3D reconstruction and robotic perception. It is challenging partially due ...
متن کاملSemantic Segmentation of Street-Side Images
In this paper we propose a method for semantic segmentation of street-side images. Segmentation and classification is pixel based and results in classes of building facades, sections of sky, road and other areas present in general images taken in the urban environment. A segmentation method is suggested and detected segments are classified. Final classification is reinforced using context infor...
متن کاملDiscriminative Fields for Modeling Spatial Dependencies in Natural Images
In this paper we present Discriminative Random Fields (DRF), a discriminative framework for the classification of natural image regions by incorporating neighborhood spatial dependencies in the labels as well as the observed data. The proposed model exploits local discriminative models and allows to relax the assumption of conditional independence of the observed data given the labels, commonly...
متن کاملCommon Landmark Discovery in Urban Scenes
In this paper, we introduce a method for unsupervised discovery of landmark objects in urban scenes. Unlike existing methods that deal with pre-defined and typically popular landmarks whose models or training data are available on the internet (e.g. Nortre Dame de Paris, Leaning tower of Pisa, etc.), we develop a new unified framework, common landmark discovery (CLD), which discovers even a pri...
متن کامل