Multi-scale Series Contextual Model for Image Parsing
نویسندگان
چکیده
Contextual information plays an important role in solving high-level vision problems and has been used widely in the field. However, using contextual information in an effective way remains a difficult problem. To address this challenge, we propose a novel framework which utilizes context information in a multi-scale structure for learning discriminative models. We apply a series of linear filters to the context image consecutively to create a scale space representation. The main idea is to take the advantage of the context image at different scales instead of a single scale giving the classifier access to a larger contextual area. Moreover, finest scale context information can be noisy while a scale space structure is more robust against noise, so our proposed method improves robustness as well as accuracy. In this framework, the improvements in accuracy between consecutive classifiers in a series architecture are larger and convergence is faster. Our strategy is general and independent of the classifier type. In other words, it has the potential to be used in any context based framework. We demonstrate performance of the algorithm on two challenging visual recognition tasks: image parsing and texture segmentation. With nearly same computational complexity our model outperforms the state of the art algorithms. Multi-scale Series Contextual Model for Image Parsing Mojtaba Seyedhosseini1,2, António R. C. Paiva1 and Tolga Tasdizen1,2 1 Scientific Computing and Imaging Institute, 2 Dept. of Electrical and Computer Eng., University of Utah, Salt Lake City, UT 84112 email: {mseyed,arpaiva,tolga}@sci.utah.edu
منابع مشابه
Detection of Neuron Membranes in Electron Microscopy Images Using Multi-scale Context and Radon-Like Features
Automated neural circuit reconstruction through electron microscopy (EM) images is a challenging problem. In this paper, we present a novel method that exploits multi-scale contextual information together with Radon-like features (RLF) to learn a series of discriminative models. The main idea is to build a framework which is capable of extracting information about cell membranes from a large co...
متن کاملLooking at Outfit to Parse Clothing
This paper extends fully-convolutional neural networks (FCN) for the clothing parsing problem. Clothing parsing requires higher-level knowledge on clothing semantics and contextual cues to disambiguate fine-grained categories. We extend FCN architecture with a side-branch network which we refer outfit encoder to predict a consistent set of clothing labels to encourage combinatorial preference, ...
متن کاملAutomatic non-parametric image parsing via hierarchical semantic voting based on sparse-dense reconstruction and spatial-contextual cues
Image parsing is vital for many high-level image understanding tasks. Although both parametric and non-parametric approaches have achieved remarkable success, many technical challenges still prevail for images containing things/objects with broad-coverage and high-variability, because it still lacks versatile and effective strategies to seamlessly integrate local–global features selection, cont...
متن کاملAn improved joint model: POS tagging and dependency parsing
Dependency parsing is a way of syntactic parsing and a natural language that automatically analyzes the dependency structure of sentences, and the input for each sentence creates a dependency graph. Part-Of-Speech (POS) tagging is a prerequisite for dependency parsing. Generally, dependency parsers do the POS tagging task along with dependency parsing in a pipeline mode. Unfortunately, in pipel...
متن کامل9 Image Segmentation by Autoregressive Time Series Model
The objective of the image segmentation is to simplify the representation of pictures into meaningful information by partitioning into image regions. Image segmentation is a software technique to locate certain objects or boundaries within an image. There are many algorithms and techniques have been developed to solve image segmentation problems for the past 20 years, though, none of the method...
متن کامل