Sampling Strategies for Bag-of-Features Image Classification
Open accessThe DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“First, it is widely accepted that denser sampling of subsequences yields a better performance [24,39], so l is set to 1.”
From Comparison of Feature Learning Methods for Human Activity Recognition Using Wearable Sensors (2018) · cites this paper as [39]
Citations
Cites
Distinctive Image Features from Scale-Invariant Keypoints
Lowe, D. G. (2004) · International Journal of Computer Vision · doi:10.1023/b:visi.0000029664.99615.94 · Cited by 55,589
Text categorization with Support Vector Machines: Learning with many relevant features
Joachims, T. (1998) · Lecture Notes in Computer Science · doi:10.1007/bfb0026683 · Cited by 7,999
Video Google: a text retrieval approach to object matching in videos
Sivic, & Zisserman (2003) · Proceedings Ninth IEEE International Conference on Computer Vision · doi:10.1109/iccv.2003.1238663 · Cited by 6,367 · cited 4×
“) as samplers [4, 1, 2, 5, 6, 7, 8, 9, 10, 11], but although such detectors have proven their value in matching applications, they were not designed to find the most informative patches for image classification and there is some evidence that they do not do so [12, 13].”
InfluentialThe Earth Mover's Distance as a Metric for Image Retrieval
Rubner, Y., Tomasi, C., & Guibas, L. J. (2000) · International Journal of Computer Vision · doi:10.1023/a:1026543900054 · Cited by 4,568
Visual categorization with bags of keypoints
Csurka, G. (2004) · Cited by 4,379
A comparison of affine region detectors
Mikolajczyk, K., Tuytelaars, T., Schmid, C., Zisserman, A., Matas, J., Schaffalitzky, F., Kadir, T., & Gool, L. V. (2005) · International Journal of Computer Vision · doi:10.1007/s11263-005-3848-x · Cited by 2,939
Object class recognition by unsupervised scale-invariant learning
Fergus, R., Perona, P., & Zisserman, A. (2003) · 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2003. Proceedings. · doi:10.1109/cvpr.2003.1211479 · Cited by 2,023
Zhang, J., Marszałek, M., Lazebnik, S., & Schmid, C. (2007) · International Journal of Computer Vision · doi:10.1007/s11263-006-9794-4 · Cited by 1,923
QBIC project: querying images by content, using color, texture, and shape:
Niblack, C. W., Barber, R., Equitz, W., Flickner, M. D., Glasman, E. H., Petkovic, D., Yanker, P., Faloutsos, C., & Taubin, G. (1993) · SPIE Proceedings · doi:10.1117/12.143648 · Cited by 1,747
Representing and Recognizing the Visual Appearance of Materials Using Three-Dimensional Textons
Leung, T., & Malik, J. (2001) · International Journal of Computer Vision · doi:10.1023/a:1011126920638 · Cited by 1,498 · cited 4×
“Patches are selected using keypoints [4, 1, 2, 5, 6, 7, 8, 9, 10, 11] or densely [3, 13, 15].”
Influential
Research works printed but not linked: 1
- Niblack, W., Barber, R., Equitz, W., Flickner, M., Glasman, D., Petkovic, D., Yanker, P.: The qbic project: Querying image by content using color, texture, and shape. SPIE 1908, 173–187 (1993)
Cited by
Items 101–200 of 925
DART: Distribution Aware Retinal Transform for Event-based Cameras
Ramesh, B., Yang, H., Orchard, G., Thi, N. A. L., Zhang, S., & Xiang, C. (2019) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2919301 · Cited by 135
“As expected, higher codebook sizes lead to better classification accuracy [51] and the standard three level SPM grid representation performs better than the two level representation.”
Building Hierarchical Representations for Oracle Character and Sketch Recognition
Guo, J. F., Wang, C., Román-Rangel, E., Chao, H., & Rui, Y. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2500019 · Cited by 106
“Moreover, Fixed-Scale-and-Dense-Sampling ScaleInvariant Feature Transform (FD-SIFT) [5], a single-scale version of the dense SIFT representation [14], was proposed”
BackgroundWheat grain classification by using dense SIFT features with SVM classifier
Olgun, M., Onarcan, A. O., Özkan, K., Işık, Ş., Sezer, O., Özgişi, K., Ayter, N. G., Başçiftçi, Z. B., Ardıç, M., & Koyuncu, O. (2016) · Computers and Electronics in Agriculture · doi:10.1016/j.compag.2016.01.033 · Cited by 101
“, 2007), object (Nowak et al., 2006) and text classification (Joachims, 1997), text representation for sentence selection (Caropreso and Matwin, 2006) and topic modeling tasks (Wallach, 2006).”
BackgroundMethodA Multilevel Point-Cluster-Based Discriminative Feature for ALS Point Cloud Classification
Zhang, Z., Zhang, L., Tong, X., Mathiopoulos, P. T., Guo, B., Huang, X., Wang, Z., & Wang, Y. (2016) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2016.2514508 · Cited by 96
“To efficiently handle the key points that are salient image patches, a bag-of-words (BoW) method [38], [39] is often used to quantize each extracted key point into a visual word and then to represent each image using a histogram of the visual words.”
MethodHarmony Potentials: Fusing Global and Local Scale for Semantic Image Segmentation
Boix, X., Gonfaus, J. M., van de Weijer, J., Bagdanov, A. D., Serrat, J., & Gonzàlez, J. (2011) · International Journal of Computer Vision · doi:10.1007/s11263-011-0449-8 · Cited by 104
“Early works incorporating topdown information include (Mori et al, 2004) which combine segmentation and recognition, and the work on image parsing pioneered by the early work of Tu and Zhu (2002) and continuing with (Tu et al, 2005; Zhu et al, 2008).”
BackgroundMethodScene Recognition on the Semantic Manifold
Kwitt, R., Vasconcelos, N. M., & Rasiwasia, N. (2012) · Lecture notes in computer science · doi:10.1007/978-3-642-33765-9_26 · Cited by 96
“In BoW methods, where recognition performance tends to increase with codebook size [17, 18], codebook generation quickly becomes a computational bottleneck.”
MethodExploring Bag of Words Architectures in the Facial Expression Domain
Sikka, K., Wu, T., Susskind, J., & Bartlett, M. S. (2012) · Lecture notes in computer science · doi:10.1007/978-3-642-33868-7_25 · Cited by 101
“Work in [18] pointed out that, (1) interest points based methods often saturate in performance since they can only provide limited patches, and (2) the minimumscale of patches has a considerable influence on results because the vast majority of patches or interest points typically occur at the finest few scales.”
BackgroundMethodA comparison of color features for visual concept classification
van de Sande, K. E. A., Gevers, T., & Snoek, C. G. M. (2008) · ACM International Conference on Image and Video Retrieval (CIVR) · doi:10.1145/1386352.1386376 · Cited by 91
“A more significant aspect of bag-of-features methods, however, is the codebook size [22].”
BackgroundSpatial Pooling of Heterogeneous Features for Image Classification
Xie, L., Tian, Q. C., Wang, M., & Zhang, B. (2014) · IEEE Transactions on Image Processing · doi:10.1109/tip.2014.2310117 · Cited by 95
Hassner, T., Mayzels, V., & Zelnik‐Manor, L. (2012) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2012.6247842 · Cited by 98
Visual Similarity Based 3D Shape Retrieval Using Bag-of-Features
Lian, Z., Godil, A., & Sun, X. (2010) · IEEE International Conference on Shape Modeling and Applications · doi:10.1109/smi.2010.20 · Cited by 90
3D Model based Object Class Detection in An Arbitrary View
Yan, P., Khan, S. M., & Shah, M. A. (2007) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2007.4409042 · Cited by 88
“One path attempts to use increasing number of local features by applying multiple feature detectors simultaneously [1, 6, 13, 14, 15].”
BackgroundHawkes processes for events in social media
Rizoiu, M.-A., Lee, Y., Mishra, S., & Xie, L. (2017) · doi:10.1145/3122865.3122874 · Cited by 128
Scene classification with low-dimensional semantic spaces and weak supervision
Rasiwasia, N., & Vasconcelos, N. M. (2008) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2008.4587372 · Cited by 86
“On one hand, it is noticed that increasing the size of the codebook improves classification performance[ 13 ].”
BackgroundPinz, A. J. (2005) · now publishers, Inc. eBooks · doi:10.1561/0600000003 · Cited by 98
Moxley, E., Kleban, J., & Manjunath, B. S. (2008) · Multimedia Information Retrieval · doi:10.1145/1460096.1460102 · Cited by 85
“SIFT Signature 11,111-dimensional, based on vocabulary tree established by Nister and Stewenius[9] using Lowe’s SIFT signature[6] extracted from 5000 random keypoints[10].”
MethodData-driven grasping with partial sensor data
Goldfeder, C., Ciocarlie, M., Peretzman, J., Dang, H., & Allen, P. K. (2009) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2009.5354078 · Cited by 84
“Unlike [27] we use a binary bag of features rather than a histogram, using the simple absence/presence rule that was shown to perform well in [26].”
BackgroundMethodMehmood, Z., Mahmood, T., & Javid, M. A. (2017) · Applied Intelligence · doi:10.1007/s10489-017-0957-5 · Cited by 89
“According to [35, 39, 43], increasing the size of the dictionary at some certain level increases the performance of image retrieval, which also results in a reduction of the semantic gap, while larger size of dictionary tends to overfit.”
BackgroundMulti-Instance Learning with Discriminative Bag Mapping
Wu, J. X., Pan, S., Xingquan Zhu, A., Chengqi Zhang, B., & Wu, X. (2018) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2017.2788430 · Cited by 84
“Vocabulary-based approaches comprise four major groups: 1) histogram-based methods use a function that maps each bag into a histogram where each bin counts how many instances fall into a specific class of the vocabulary [13], [35], [36], [37]; 2) distancebased methodsmap each bag into a vector space by providing the l…”
MethodA general framework for sensor-based human activity recognition
Köping, L., Shirahama, K., & Grzegorzek, M. (2018) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2017.12.025 · Cited by 93
Discriminative cue integration for medical image annotation
Tommasi, T., Orabona, F., & Barbara Caputo, B. (2008) · Pattern Recognition Letters · doi:10.1016/j.patrec.2008.03.009 · Cited by 82
“Indeed, psychophysical evidence suggest that natural vision-based classification tasks are performed better when multiple visual cues can be combined to reduce ambiguity [Tanaka et al. (1991)].”
InfluentialImage retrieval by addition of spatial information based on histograms of triangular regions
Ali, N., Bajwa, K. B., Sablatnig, R., & Mehmood, Z. (2016) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2016.04.002 · Cited by 82
Miyaki, R., Yoshida, S., Tanaka, S., Kominami, Y., Sanomura, Y., Matsuo, T., Oka, S., Raytchev, B., Tamaki, T., Koide, T., et al. (2014) · Journal of Clinical Gastroenterology · doi:10.1097/mcg.0000000000000104 · Cited by 81
Visual Synset: Towards a higher-level visual representation
Zheng, Y.-T., Zhao, M., Neo, S.-Y., Chua, T., & Tian, Q. (2008) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2008.4587611 · Cited by 79
“The visual synset delivers a comparable result with the state-of-arts system [14], which however extracts quite a large amount (10k) of regions per image for classification.”
Web-scale computer vision using MapReduce for multimedia data mining
White, B., YEH, T. P., Lin, J., & Davis, L. S. (2010) · International Workshop on Multimedia Data Mining (MDM/KDD) · doi:10.1145/1814245.1814254 · Cited by 77
“To compute BoF vectors, local feature points are selected by a detection algorithm [3] or randomly [16], the local features are clustered, and a histogram is calculated from the local feature quantizations.”
MethodCrowdsourcing image analysis for plant phenomics to generate ground truth data for machine learning
Zhou, N., Siegel, Z. D., Zarecor, S., Lee, N., Campbell, D. A., Andorf, C. M., Nettleton, D., Lawrence-Dill, C. J., Ganapathysubramanian, B., Kelly, J. W., et al. (2018) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1006337 · Cited by 76
Spatio-temporal context analysis within video volumes for anomalous-event detection and localization
Li, N., Wu, X., Xu, D., Guo, H., & Feng, W. (2015) · Neurocomputing · doi:10.1016/j.neucom.2014.12.064 · Cited by 75
“We densely sample the video to obtain spatio-temporal features which describe both the appearance and motion of events, and then apply an improved BOW [43–45] method to represent the spatio-temporal configuration of video volumes within large video cubes.”
MethodAutomatic detection of child pornography using color visual words
Ulges, A., & Stahl, A. (2011) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme.2011.6011977 · Cited by 85
“The resulting patches are vector quantized using a visual codebook of prototypes learned by a K-means clustering over a large-scale set of training images (we use K = 2, 000, which can be considered a common choice (29)).”
MethodJohnny: An Autonomous Service Robot for Domestic Environments
Breuer, T., Macedo, G. R. G., Hartanto, R., Hochgeschwender, N., Holz, D., Hegger, F., Jin, Z., Müller, C. A., Paulus, J., Reckhaus, M., et al. (2011) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-011-9608-y · Cited by 80
“The system is grounded on 2D image information and relies on a geometric-free approach called Bag of Features(BoF) [17, 39, 55].”
BackgroundMethodSpatiotemporal bag-of-features for early wildfire smoke detection
Ko, B. C., Park, J., & Nam, J.-Y. (2013) · Image and Vision Computing · doi:10.1016/j.imavis.2013.08.001 · Cited by 86
Multiscale fully convolutional network with application to industrial inspection
Bian, X., Lim, S. N., & Zhou, N. (2016) · 2016 IEEE Winter Conference on Applications of Computer Vision (WACV) · doi:10.1109/wacv.2016.7477595 · Cited by 72
“Different from typical two-stage approaches that involve hand-crafted feature extraction and classifier design [12, 17], CNN automatically learns the most informative features from the input images themselves for the task involved by using a stack of convolutional layers [8].”
MethodBag of spatio-visual words for context inference in scene classification
Bolovinou, A., Pratikakis, I., & Perantonis, S. J. (2012) · Pattern Recognition · doi:10.1016/j.patcog.2012.07.024 · Cited by 79
Nayak, A., Chakraborty, S., & Swain, D. K. (2023) · Smart Agricultural Technology · doi:10.1016/j.atech.2023.100195 · Cited by 97
Multi-view action recognition using local similarity random forests and sensor fusion
Zhu, F., Shao, L., & Lin, M. (2012) · Pattern Recognition Letters · doi:10.1016/j.patrec.2012.04.016 · Cited by 76
“To demonstrate the effectiveness of our method, we first compare the results of our method with those of the baseline BoW (Nowak et al., 2006) and the NBNN (Boiman et al.”
ResultInfluentialFace recognition using the POEM descriptor
Vu, N., Dee, H., & Caplier, A. (2012) · Pattern Recognition · doi:10.1016/j.patcog.2011.12.021 · Cited by 71
Super Fast Event Recognition in Internet Videos
Jiang, Y., Dai, Q., Mei, T., Rui, Y., & Chang, S. (2015) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2015.2436813 · Cited by 74
Liver disease classification from ultrasound using multi-scale CNN
Che, H., Brown, L. G., Foran, D. J., Nosher, J. L., & Hacihaliloglu, I. (2021) · International Journal of Computer Assisted Radiology and Surgery · doi:10.1007/s11548-021-02414-0 · Cited by 78
“…our network architectures were compared with each other and two popular approaches: (1) Features are extracted using bag-of-features (BOF) method [25],and SVM is used to perform the classification; (2) a pre-trained CNN is utilized to extract features with the transfer learning concept, and…”
BackgroundMethodMalignant melanoma detection by Bag-of-Features classification
Situ, N., Yuan, X., Chen, J., & Zouridakis, G. (2008) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/iembs.2008.4649862 · Cited by 68
“Bag-of-Features based image classification is widely used in computer vision [3], [4].”
BackgroundContext-Based Object-Class Recognition and Retrieval by Generalized Correlograms
Amores, J., Sebe, N., & Radeva, P. I. (2007) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2007.1098 · Cited by 67
Stanciu, S. G., Xu, S., Peng, Q., Yan, J., Stanciu, G. A., Welsch, R. E., So, P. T. C., Csúcs, G., & Yu, H. (2014) · Scientific Reports · doi:10.1038/srep04636 · Cited by 67
Image Reconstruction from Bag-of-Visual-Words
Kato, H., & Harada, T. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.127 · Cited by 81
“The latter, designated as dense sampling, is better for image recognition and retrieval [31], [32].”
BackgroundAssistive Clothing Pattern Recognition for Visually Impaired People
Yang, X., Yuan, S., & Tian, Y. (2014) · IEEE Transactions on Human-Machine Systems · doi:10.1109/thms.2014.2302814 · Cited by 64
“Process of local image feature extraction. paper, the uniform grids are used as the interest points sampling strategy, as more sophisticated detectors tend to saturate and fail to provide enough interest points, especially for the textureless images [18].”
BackgroundMethodInfluentialNon-empirical identification of trigger sites in heterogeneous processes using persistent homology
Kimura, M., Obayashi, I., Takeichi, Y., Murao, R., & Hiraoka, Y. (2018) · Scientific Reports · doi:10.1038/s41598-018-21867-z · Cited by 69
Boundary preserving dense local regions
Kim, J., & Grauman, K. L. (2011) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2011.5995526 · Cited by 69
“Researchers have developed a variety of techniques to perform detection, ranging from sophisticated interest point operators [19, 20, 21, 12, 31] to dense sampling strategies [23].”
Background3D object recognition and classification: a systematic literature review
Carvalho, L. E., & von Wangenheim, A. (2019) · Pattern Analysis and Applications · doi:10.1007/s10044-019-00804-4 · Cited by 66
“Based on this framework, the performance of five codebook models was compared using a variety of sampling strategies combinations (sparse via the SIFT 3D point of interest detector, and dense as recommended in [214]), feature encoding techniques (k-means clustering and extremely randomized clustering (ERC) forests) an…”
MethodDeep convolutional neural networks for annotating gene expression patterns in the mouse brain
Zeng, T., Li, R., Mukkamala, R., Ye, J., & Ji, S. (2015) · BMC Bioinformatics · doi:10.1186/s12859-015-0553-9 · Cited by 61
“We first considered a baseline approach based on the bag-of-words representation that has been widely used in modeling natural and biological images [5,15-20].”
BackgroundCharacterizing powder materials using keypoint-based computer vision methods
DeCost, B., & Holm, E. A. (2016) · Computational Materials Science · doi:10.1016/j.commatsci.2016.08.038 · Cited by 65
Miyaki, R., Yoshida, S., Tanaka, S., Kominami, Y., Sanomura, Y., Matsuo, T., Oka, S., Raytchev, B., Tamaki, T., Koide, T., et al. (2013) · Journal of Gastroenterology and Hepatology · doi:10.1111/jgh.12149 · Cited by 63
A review of automated image understanding within 3D baggage computed tomography security screening
Mouton, A., & Breckon, T. P. (2015) · Journal of X-Ray Science and Technology · doi:10.3233/xst-150508 · Cited by 63
“[105] demonstrates an improvement over the 3D visual cortex model in terms of classification accuracy and processing time using a codebook approach based on Extremely Randomised Clustering (ERC) forests [106], a dense feature sampling strategy [107] and an SVM classifier [69].”
Classification of Endomicroscopic Images of the Lung Based on Random Subwindows and Extra-Trees
Désir, C., Petitjean, C., Heutte, L., Salaün, M., & Thiberville, L. (2012) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2012.2204747 · Cited by 61
A Hybrid Geometric Spatial Image Representation for scene classification
Ali, N., Zafar, B., Riaz, F., Dar, S. H., Ratyal, N. I., Bajwa, K. B., Iqbal, M. K., & Sajid, M. (2018) · PLoS ONE · doi:10.1371/journal.pone.0203339 · Cited by 65
e-TLD: Event-Based Framework for Dynamic Object Tracking
Ramesh, B., Zhang, S., Yang, H., Ussa, A., Ong, M., Orchard, G., & Xiang, C. (2020) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2020.3044287 · Cited by 68
A Model of Visual Attention for Natural Image Retrieval
Liu, G., & Fan, D.-P. (2013) · International Conference on Information Science and Cloud Computing Companion (ISCC-C) · doi:10.1109/iscc-c.2013.21 · Cited by 60
Global Gaussian approach for scene categorization using information geometry
Nakayama, H., Harada, T., & Kuniyoshi, Y. (2010) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2010.5539921 · Cited by 57
“As the feature sampling method, we take the dense sampling strategy, following many successful works related to image categorization [9, 20, 6, 29, 30].”
MethodClassifying Compound Structures in Satellite Images: A Compressed Representation for Fast Queries
Gueguen, L. (2014) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2014.2348864 · Cited by 57
“It is well known that the number of visual words that compose the dictionary influences the expected classification accuracies [2].”
BackgroundVisual word proximity and linguistics for semantic video indexing and near-duplicate retrieval
Jiang, Y., & Ngo, C. (2008) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2008.10.002 · Cited by 57
“For example, TF-IDF is adopted in video google [17], binary weighting is employed in [1] for image classification, and most approaches use TF directly [2,3].”
Gueguen, L., & Hamid, R. (2016) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2016.2516402 · Cited by 55
“In recent years, the bag-of-visual-words model has been extremely popular in computer vision for categorizing large corpus of images [26], [27].”
BackgroundA Comparison of Dense Region Detectors for Image Search and Fine-Grained Classification
İşcen, A., Tolias, G., Gosselin, P.-H., & Jeǵou, H. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2423557 · Cited by 56
Recent Advances in Baggage Threat Detection: A Comprehensive and Systematic Survey
Velayudhan, D., Hassan, T., Damiani, E., & Werghi, N. (2022) · ACM Computing Surveys · doi:10.1145/3549932 · Cited by 65
“[146] combined diferent sampling techniques (3D SIFT based sparse and dense sampling [156]), feature encoding methods (based on k-means [105] and Extremely Randomised Clustering forests (ERC) [141]) and classiiers (SVM [35] and Random forests [26]) to study the performance of object classiication in pre-segmented volu…”
MethodMultimodal blending for high-accuracy instance recognition
Xie, Z., Singh, A., Uang, J., Narayan, K. S., & Abbeel, P. (2013) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2013.6696666 · Cited by 54
“also illustrate the importance of dense sampling for bag-of-features image classification [12].”
BackgroundCM-BOF: visual similarity-based 3D shape retrieval using Clock Matching and Bag-of-Features
Lian, Z., Godil, A., Sun, X., & Xiao, J. (2013) · Machine Vision and Applications · doi:10.1007/s00138-013-0501-5 · Cited by 53
“The previous work [36] has also demonstrated that the first codebook construction method results in better performance in image classification against the second method.”
MethodFAST MULTI-CLASS IMAGE ANNOTATION WITH RANDOM SUBWINDOWS AND MULTIPLE OUTPUT RANDOMIZED TREES
Dumont, M., Marée, R., Wehenkel, L. A., & Geurts, P. (2009) · International Conference on Computer Vision Theory and Applications · doi:10.5220/0001800001960203 · Cited by 52
“Randomization of patch extraction was first proposed in (Marée et al., 2003) for image classification and subsequently in (Marée et al., 2005; Nowak et al., 2006).”
MethodFeature-based brain MRI retrieval for Alzheimer disease diagnosis
Mizotin, M. M., Benois‐Pineau, J., Michèle Allard, A., & Catheline, G. (2012) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2012.6467091 · Cited by 53
“No significant dependence of the results on the sampling strategy has been found, in accordance with the work [12].”
ResultWhat is the spatial extent of an object?
Uijlings, J. R. R., Smeulders, A. W. M., & Scha, R. J. H. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206663 · Cited by 51
“The popular Bag of Words framework produced the best results in two image/video retrieval tasks of 2008: The Pascal VOC challenge [4] and the TrecVid Video Retrieval task [18].”
MethodJoutsijoki, H., Haponen, M., Rasku, J., Aalto‐Setälä, K., & Juhola, M. (2016) · Computational and Mathematical Methods in Medicine · doi:10.1155/2016/3091039 · Cited by 56
Dense Segmentation-Aware Descriptors
Trulls, E., Kokkinos, I., Sanfeliu, A., & Moreno-Noguer, F. (2013) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2013.372 · Cited by 53
“A different thread of works, such as Daisy [32], dense SIFT [34] or the dense Scale-Invariant Descriptors [13] have demonstrated that it is possible to efficiently compute descriptors densely , i.e. for every pixel, and use them as a generic low-level image representation on a par with filterbanks.”
Asymmetric region-to-image matching for comparing images with generic object categories
Kim, J., & Grauman, K. L. (2010) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2010.5539923 · Cited by 54
“While generally effective for matching object instances, sparsely sampled interest points provide a weaker representation for category-level matching; much evidence suggests that a dense coverage of features is preferable [8].”
BackgroundA temporal Bayesian model for classifying, detecting and localizing activities in video sequences
Malgireddy, M. R., Nwogu, I., & Govindaraju, V. (2012) · 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops · doi:10.1109/cvprw.2012.6239185 · Cited by 51
“Similar results were also shown by [18] for image classification.”
ResultShape classification using invariant features and contextual information in the bag-of-words model
Ramesh, B., Xiang, C., & Lee, T. H. (2014) · Pattern Recognition · doi:10.1016/j.patcog.2014.09.019 · Cited by 51
“In general, higher classification accuracy has been linked to a larger vocabulary [28,33], but saturation can be expected at some point [33].”
BackgroundImproving Deep Forest by Confidence Screening
Pang, M., Ting, K. M., Zhao, P., & Zhou, Z. (2018) · Proceedings · doi:10.1109/icdm.2018.00158 · Cited by 60
“The subsampling strategy is related to sampling strategies for bag-of-features image classification [14], [15].”
MethodPattern of Local Gravitational Force (PLGF): A Novel Local Image Descriptor
Bhattacharjee, D., & Roy, H. (2019) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2930192 · Cited by 56
“[16] gave stronger arguments in favor of dense descriptors.”
BackgroundSketch-Based Image Retrieval Through Hypothesis-Driven Object Boundary Selection With HLR Descriptor
Wang, S., Zhang, J., Han, T. X., & Miao, Z. (2015) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2015.2431492 · Cited by 50
“[35] found that (1) dense sampling captures the maximum information at the cost of higher computational load, and (2) key-point sampling may result in the loss of important information.”
BackgroundLog‐Euclidean bag of words for human action recognition
Faraki, M., Palhang, M., & Sanderson, C. (2014) · IET Computer Vision · doi:10.1049/iet-cvi.2014.0018 · Cited by 52
“Within the fields of image categorisation and face recognition it has been shown that discrimination approaches based on Bag of Words (BoW) are effective [29, 36, 49, 51].”
Bioimage Classification with Handcrafted and Learned Features
Nanni, L., Brahnam, S., Ghidoni, S., & Lumini, A. (2018) · IEEE Transactions on Computational Biology and Bioinformatics · doi:10.1109/tcbb.2018.2821127 · Cited by 53
“CLM, unlike BoF [59], uses a”
Object classification in 3D baggage security computed tomography imagery using visual codebooks
Flitton, G. T., Mouton, A., & Breckon, T. P. (2015) · Pattern Recognition · doi:10.1016/j.patcog.2015.02.006 · Cited by 51
“Traditionally, the BoW model is composed of the following steps: [51]: 1) feature detection and description (Section 2 and 3); 2) visual codebook generation and 3) classification.”
MethodZhu, L., Klein, D. A., Frintrop, S., Cao, Z., & Cremers, A. B. (2014) · IEEE Transactions on Image Processing · doi:10.1109/tip.2014.2361024 · Cited by 49
A fast and efficient image retrieval system based on color and texture features
Singh, C., & Kaur, K. P. (2016) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2016.10.002 · Cited by 52
Learning and Recognition of On-Premise Signs From Weakly Labeled Street View Images
Tsai, T.-H., Cheng, W.-H., You, C.-W., Hu, M., Tsui, A. W., & Chi, H.-Y. (2014) · IEEE Transactions on Image Processing · doi:10.1109/tip.2014.2298982 · Cited by 48
Image classification by addition of spatial information based on histograms of orthogonal vectors
Zafar, B., Ashraf, R., Ali, N., Ahmed, M., Jabbar, S., & Chatzichristofis, S. A. (2018) · PLoS ONE · doi:10.1371/journal.pone.0198175 · Cited by 47
Houwen, B. B. S. L., Hazewinkel, Y., Giotis, I., Vleugels, J. L. A., Mostafavi, N. S., van Putten, P., Fockens, P., Dekker, E., & POLAR Study Group (2023) · Endoscopy · doi:10.1055/a-2009-3990 · Cited by 48
“The classifier that consistently outperformed all other models in a fivefold cross validation setting was finally based on a Bag of Visual Words approach using SIFT features that are extracted from each sub-image by means of grid sampling (gridSIFT) [19, 20].”
MethodActive SLAM using 3D Submap Saliency for Underwater Volumetric Exploration
Suresh, S., Sodhi, P., Mangelson, J. G., Wettergreen, D. S., & Kaess, M. (2020) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra40945.2020.9196939 · Cited by 56
“A bag-of-words (BoW) representation [26] creates a scene dictionary, which can be used for online loop closure detection [27].”
BackgroundScene Recognition with Camera Phones for Tourist Information Access
Lim, J., Li, Y., You, Y., & Chevallet, J. (2007) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme.2007.4284596 · Cited by 47
“We adopt multi-scale uniform sampling to extract patches from images instead of interest point scheme as the latter does not have advantage over random and uniform samplings when the sampling is dense enough [ 8 ].”
MethodImage Retrieval based on Bag-of-Words modelPreprint
Liu, J. (2013) · arXiv (Cornell University) · doi:10.48550/arxiv.1304.5168 · Cited by 56
“It is noted that the BoW approach, although is simple and directly borrowed from text retrieval community, has shown excellent performance not only for CBIR task but also for other vision tasks like object recognition, image classification and annotation [18, 15].”
A study of Bag-of-Visual-Words representations for handwritten keyword spotting
Aldavert, D., Rusiñol, M., Toledo, R., & Lladós, J. (2015) · International Journal on Document Analysis and Recognition (IJDAR) · doi:10.1007/s10032-015-0245-z · Cited by 56
“demonstrate in [32] that the larger the number of regions, the better the results.”
ResultTeixeira, L. F., & Côrte-Real, L. (2008) · Pattern Recognition Letters · doi:10.1016/j.patrec.2008.04.001 · Cited by 46
“We used, therefore, random point selection from a pyramid with regular grids, as suggested by Nowak et al. (2006), since this method provides a dense representation of each object image.”
BackgroundMethodInfluentialDeep Convolutional Neural Networks for Multi-instance Multi-task Learning
Zeng, T., & Ji, S. (2015) · Proceedings · doi:10.1109/icdm.2015.92 · Cited by 48
Determining Patch Saliency Using Low-Level Context
Parikh, D., Zitnick, C. L., & Chen, T. (2008) · Lecture notes in computer science · doi:10.1007/978-3-540-88688-4_33 · Cited by 46
“In contrast to previous results that show recognition accuracy increases with the density of the sampling [7], the contextual measures achieve maximal accuracy using a sparse sampling.”
BackgroundMethodResultInfluentialAn efficient divide-and-conquer cascade for nonlinear object detection
Lampert, C. H. (2010) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2010.5540107 · Cited by 47
“Multiple publications over the last years have shown that detection performance is improved by the use of densely samples feature points and large codebooks [14], multiple sets of feature descriptors [19], spatial pyramid representations [12], and the use of image context [17].”
BackgroundFast and robust object segmentation with the Integral Linear Classifier
Aldavert, D., de Mántaras, R. L., Arnau Ramisa, A., & Toledo, R. (2010) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2010.5540098 · Cited by 42
“Using densely sampled image descriptors as input data has several advantages when compared to keypoint-based approaches: more information can be extracted from the underlying image and the time-consuming keypoint detection step is not needed [18].”
A bag-of-words approach for Drosophila gene expression pattern annotation
Ji, S., Li, Y.-X., Zhou, Z.-H., Kumar, S., & Ye, J. (2009) · BMC Bioinformatics · doi:10.1186/1471-2105-10-119 · Cited by 48
“Three common methods for generating local patches on images are those based on affine region detectors [22], random sampling [23], and regular patches [24].”
MethodUnsupervised Multiclass Region Cosegmentation via Ensemble Clustering and Energy Minimization
Li, H., Meng, F., Wu, Q., & Luo, B. (2013) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2013.2280851 · Cited by 44
“It is noted that dense sampling at every pixel can capture the most object information, but memory and computational load are very intensive [54].”
BackgroundJoint Optimization Toward Effective and Efficient Image Search
Wei, S., Xu, D., Li, X., & Zhao, Y. (2013) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2013.2245890 · Cited by 43
“As indicated in existing work [7], two key factors that affect quantization error are visual vocabulary size K and vocabulary construction method.”
BackgroundMethodLet the kernel figure it out; Principled learning of pre-processing for kernel classifiers
Gehler, P., & Nowozin, S. (2009) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvprw.2009.5206592 · Cited by 42
Interactive image segmentation using probabilistic hypergraphs
Ding, L., & Yılmaz, A. (2009) · Pattern Recognition · doi:10.1016/j.patcog.2009.11.025 · Cited by 52
“which is essentially the bag-of-features model [44] commonly used in the computer vision literature.”
MethodBuilding Recognition Using Local Oriented Features
Li, J., & Allinson, N. M. (2013) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2013.2245910 · Cited by 55
“In SIFT [30], orientations are measured in the neighborhood of a keypoint, resulting in local orientation histograms; bag-of-features [35] use pooling for vector quantization of local descriptors.”
MethodDiscriminative BoW Framework for Mobile Landmark Recognition
Chen, T., & Yap, K. (2013) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2013.2267015 · Cited by 41
“These works can be categorized into two classes of image representation: 1) local patch image representation [3]–[5], [44] that uses visual features extracted from the local patches in the image for recognition; and 2) bag-of-words (BoW) histogram representation [7]–[12], [34] that generates a BoW histogram for each i…”
BackgroundLearning class-specific affinities for image labelling
Batra, D., Sukthankar, R., & Chen, T. (2008) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2008.4587432 · Cited by 41
“Following the popular bag-of-words model [22], we discretize/quantize this feature space to get “visual words”.”
MethodZhan, J., Ge, X. J., Huang, S., Zhao, L., Wong, J. K. W., & He, X. (2019) · Facilities · doi:10.1108/f-01-2018-0005 · Cited by 42
“Dense sampling was used to extract features in this system, by splitting images with “grip”, which shows suitability in indoor environments (Nowak et al. 2006).”
MethodLearning a mixture of sparse distance metrics for classification and dimensionality reduction
Hong, Y., Li, Q., Jiang, J., & Zhuowen Tu, A. (2011) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2011.6126332 · Cited by 40
D-Nets: Beyond patch-based image descriptors
Hundelshausen, F. V., & Sukthankar, R. (2012) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2012.6248022 · Cited by 40
“Due to space limits, we restrict ourselves to three key observations: • Unlike patch-based descriptors, which derive their scale and rotation invariance from the interest point detector (and lose these under dense sampling [13]), the dense variant of D-Nets retains all of its original invariance properties.”
Method
Metadata
- Authors
- Nowak, E., Jurie, F., & Triggs, B.
- Affiliations
- Centre National de la Recherche Scientifique (FR)
- Venue
- Lecture Notes in Computer Science
- Publisher
- Springer Berlin Heidelberg
- Published
- 2006
- Volume
- 3954
- Pages
- 490-503
- ISSN
- 0302-9743 (ISSN-L)
- 1611-3349
- OpenAlex
- W2171896402
- Cited by
- 927
- References
- 26
- Citations by year
- 2026: 1 · 2025: 3 · 2024: 7 · 2023: 10 · 2022: 11 · 2021: 17 · 2020: 35 · 2019: 44 · 2018: 55 · 2017: 59 · 2016: 73 · 2015: 77 · 2014: 78 · 2013: 82 · 2012: 56
- FWCI
- 42.67
- Type
- conference paper
- Language
- en
- Open access
- green
- hal.science
- Licence
- other-oa
- www.springer.com/tdm (tdm)
- Locations
- Lecture notes in computer science (publishedVersion)
- HAL (submittedVersion, other-oa)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Image Retrieval and Classification Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Image and Video Retrieval Techniques
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Radiomics and Machine Learning in Medical Imaging
- Keywords
- bag of visual words, content-based image classification, patch sampling, random sampling, image descriptors, codebook size, histogram normalization, texture analysis
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.