Bag of spatio-visual words for context inference in scene classification

Bolovinou, Anastasia; Pratikakis, Ioannis; Perantonis, Stavros

doi:10.1016/j.patcog.2012.07.024

Cited by 67 publications

(47 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the domain of object class recognition, the Bag of (Visual) Words (BoW), or codebook, approach has been met with considerable success [25,[41][42][43]. In terms of object recognition in 3D baggage-CT imagery, Flitton et al [8] compare the performance of a 3D visual cortexbased approach to a BoW model using the 3D SIFT descriptor [6].…”

Section: Introductionmentioning

confidence: 99%

Object classification in 3D baggage security computed tomography imagery using visual codebooks

Flitton

Mouton

Breckon

2015

Pattern Recognition

View full text Add to dashboard Cite

Use policyThe full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-pro t purposes provided that:• a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.Please consult the full DRO policy for further details. AbstractWe investigate the performance of a Bag of (Visual) Words ( Optimal performance is achieved using the DH and DGH descriptors in conjunction with an uncertainty assignment methodology. Successful detection rates in excess of 97% for handguns and 89% for bottles and false-positive rates of approximately 2-3% are achieved. We demonstrate that the underlying imaging modality and the irrelevance of illumination and scale invariance within the transmission imagery context considered here, result in the favourable performance of simpler density histogram descriptors (DH, DGH) over 3D extensions of the well-established SIFT and RIFT feature descriptor approaches.

show abstract

Section: Introductionmentioning

confidence: 99%

Object classification in 3D baggage security computed tomography imagery using visual codebooks

Flitton

Mouton

Breckon

2015

Pattern Recognition

View full text Add to dashboard Cite

show abstract

“…Image segmentation is a classical problem of image processing, which is being studied during the last decade (1). Its main aim is to split the image into disjoint regions formed by pixels; each such region is bordering a meaningful object or area in the image.…”

Section: Image Segmentationmentioning

confidence: 99%

“…The disadvantage is a long segmentation time and the need for participation of a person in the process of segmentation (1,2,3,4,5,6). Fig.…”

Section: Active Contour Modelsmentioning

confidence: 99%

Use of graph algorithms in the processing and analysis of images with focus on the biomedical data

Ždímalová¹,

Roznovjak²,

Weismann³

et al. 2017

BLL

View full text Add to dashboard Cite

Image segmentation is a known problem in the fi eld of image processing. A great number of methods based on different approaches to this issue was created. One of these approaches utilizes the fi ndings of the graph theory. METHODS: Our work focuses on segmentation using shortest paths in a graph. Specifi cally, we deal with methods of "Intelligent Scissors," which use Dijkstra's algorithm to fi nd the shortest paths. RESULTS: We created a new software in Microsoft Visual Studio 2013 integrated development environment Visual C++ in the language C++/CLI. We created a format application with a graphical users development environment for system Windows, with using the platform .Net (version 4.5). The program was used for handling and processing the original medical data. CONCLUSION: The major disadvantage of the method of "Intelligent Scissors" is the computational time length of Dijkstra's algorithm. However, after the implementation of a more effi cient priority queue, this problem could be alleviated. The main advantage of this method we see in training that enables to adapt to a particular kind of edge, which we need to segment. The user involvement has a signifi cant infl uence on the process of segmentation, which enormously aids to achieve high-quality results (Fig. 7, Ref. 13). Text in PDF www.elis.sk.

show abstract

“…Motivated by SPM and SCK, a pyramid-of-spatial-relations model [9] introduces a novel concept to describe quantized relative relationship of a set of local features and outperforms both BOW and SCK. Bolovinou et al [35] proposed a bag-of-spatio-visual-words model (BoSVW), which incorporates local context information into the BOW representation and can efficiently tackle the problem of high-dimensional spatial feature clustering by introducing the spherical K-means algorithm. In general, all of these methods strongly rely on the extraction of the hand-crafted low-level features, learning the codebook and coding local features, which are usually highly time consuming.…”

Section: Related Workmentioning

confidence: 99%

Fast Binary Coding for the Scene Classification of High-Resolution Remote Sensing Imagery

Xia

et al. 2016

Remote Sensing

View full text Add to dashboard Cite

Scene classification of high-resolution remote sensing (HRRS) imagery is an important task in the intelligent processing of remote sensing images and has attracted much attention in recent years. Although the existing scene classification methods, e.g., the bag-of-words (BOW) model and its variants, can achieve acceptable performance, these approaches strongly rely on the extraction of local features and the complicated coding strategy, which are usually time consuming and demand much expert effort. In this paper, we propose a fast binary coding (FBC) method, to effectively generate efficient discriminative scene representations of HRRS images. The main idea is inspired by the unsupervised feature learning technique and the binary feature descriptions. More precisely, equipped with the unsupervised feature learning technique, we first learn a set of optimal "filters" from large quantities of randomly-sampled image patches and then obtain feature maps by convolving the image scene with the learned filters. After binarizing the feature maps, we perform a simple hashing step to convert the binary-valued feature map to the integer-valued feature map. Finally, statistical histograms computed on the integer-valued feature map are used as global feature representations of the scenes of HRRS images, similar to the conventional BOW model. The analysis of the algorithm complexity and experiments on HRRS image datasets demonstrate that, in contrast with existing scene classification approaches, the proposed FBC has much faster computational speed and achieves comparable classification performance. In addition, we also propose two extensions to FBC, i.e., the spatial co-occurrence matrix and different visual saliency maps, for further improving its final classification accuracy.

show abstract

Bag of spatio-visual words for context inference in scene classification

Cited by 67 publications

References 50 publications

Object classification in 3D baggage security computed tomography imagery using visual codebooks

Object classification in 3D baggage security computed tomography imagery using visual codebooks

Use of graph algorithms in the processing and analysis of images with focus on the biomedical data

Fast Binary Coding for the Scene Classification of High-Resolution Remote Sensing Imagery

Contact Info

Product

Resources

About