Ensemble of exemplar-SVMs for object detection and beyond
Open accessThe DOI matches its Crossref record; 23 of the 25 references it cites were found in a registry or an index.
Abstract
This paper proposes a conceptually simple but surprisingly powerful method which combines the effectiveness of a discriminative object detector with the explicit correspondence offered by a nearest-neighbor approach. The method is based on training a separate linear SVM classifier for every exemplar in the training set. Each of these Exemplar-SVMs is thus defined by a single positive instance and millions of negatives. While each detector is quite specific to its exemplar, we empirically observe that an ensemble of such Exemplar-SVMs offers surprisingly good generalization. Our performance on the PASCAL VOC detection task is on par with the much more complex latent part-based model of Felzenszwalb et al., at only a modest computational cost increase. But the central benefit of our approach is that it creates an explicit association between each detection and a single training exemplar. Because most detections show good alignment to their associated exemplar, it is possible to transfer any available exemplar meta-data (segmentation, geometric structure, 3D model, etc.) directly onto the detections, which can then be used as part of overall scene understanding.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures


![Figure 3 Comparison. Given a bicycle training sample from PASCAL (represented with a HOG weight vector w ), we show the top 6 matches from the PASCAL test-set using three methods. Row 1: naive nearest neighbor (using raw normalized HOG). Row 2: Trained Exemplar-SVM (notice how w focuses on bike-specific edges). Row 3: Learned distance function - an Exemplar-SVM but trained in the 'distance-to-exemplar' vector space, with the exemplar being placed at the origin (loosely corresponding to [10, 14]).](https://figures.cochat.ai/paper-figures/1cc4b956c3534a219577db00fbbb8da4/ensemble-of-exemplar-svms-for-object-detection-and-beyond-figure-3.png)







Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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
“In the case of sliding-window object detectors, such as the deformable parts model (DPM) [12], this imbalance may be as extreme as 100,000 background examples to every one object.The recent trend towards object-proposal-based detectors [15,32] mitigates this issue to an extent, but the imbalance ratio may still be high (e.g., 70:1).”
From Training Region-Based Object Detectors with Online Hard Example Mining (2016) · cites this paper as [32]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
LIBSVM: A library for support vector machines
Chang, C.-C., & Lin, C.-J. (2011) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/1961189.1961199 · Cited by 40,918 · cited 1×
“Using the procedure above, we train an ensemble of Exemplar-SVM, one for each positive instance in the training set.”
MethodHistograms of Oriented Gradients for Human Detection
Dalal, N., & Triggs, B. (2005) · 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) · doi:10.1109/cvpr.2005.177 · Cited by 33,584 · cited 2×
“Yet today, a number of methods can do just that with reasonable accuracy.”
BackgroundThe pascal visual object classes (VOC) challenge
Everingham, M., Van Gool, L., Williams, C. K. I., Winn, J., & Zisserman, A. (2010) · International Journal of Computer Vision · doi:10.1007/s11263-009-0275-4 · Cited by 19,599 · cited 2×
“For our experiments, we use a single source of exemplars: the PASCAL VOC 2007 dataset [8] – a popular dataset used to benchmark object detection algorithms.”
BackgroundMethodLibsvm : A library for support vector machines
Ferrari, V. (2008) · Medical Entomology and Zoology · Cited by 10,122
Object Detection with Discriminatively Trained Part-Based Models
Felzenszwalb, P. F., Girshick, R. B., McAllester, D., & Ramanan, D. (2010) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2009.167 · Cited by 9,853 · cited 4×
“For this, we present results on a set of tasks including segmentation, qualitative-geometry estimation, 3D model transfer, and related object priming.”
BackgroundMethodInfluentialProbabilistic Outputs for Support vector Machines and Comparisons to Regularized Likelihood Methods
Platt, J. C. (1999) · Cited by 6,384
Shape matching and object recognition using shape contexts
Belongie, S., Malik, J., & Puzicha, J. (2002) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/34.993558 · Cited by 6,306 · cited 1×
“Because most detections show good alignment to their associated exemplar, it is possible to transfer any available exemplar meta-data (segmentation, geometric structure, 3D model, etc.) directly onto the detections, which can then be used as part of overall scene understanding.”
BackgroundEstimating the Support of a High-Dimensional Distribution
Schölkopf, B., Platt, J. C., Shawe-Taylor, J., Smola, A. J., & Williamson, R. C. (2001) · Neural Computation · doi:10.1162/089976601750264965 · Cited by 6,168
SUN database: Large-scale scene recognition from abbey to zoo
Xiao, J., Hays, J., Ehinger, K. A., Oliva, A., & Torralba, A. (2010) · 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2010.5539970 · Cited by 3,283 · cited 1×
“Because most detections show good alignment to their associated exemplar, it is possible to transfer any available exemplar meta-data (segmentation, geometric structure, 3D model, etc.) directly onto the detections, which can then be used as part of overall scene understanding.”
Background80 Million Tiny Images: A Large Data Set for Nonparametric Object and Scene Recognition
Torralba, A., Fergus, R., & Freeman, W. T. (2008) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2008.128 · Cited by 1,921
Research works printed but not linked: 1
- A. Frome and J. Malik. Image retrieval and recognition using local distance functions. NIPS, 2006. 2, 3, 4
Cited by
Items 201–300 of 873
Shen, W., Bai, X., Hu, Z., & Zhang, Z. (2015) · Pattern Recognition · doi:10.1016/j.patcog.2015.10.015 · Cited by 53
“Although the multiple classifiers are trained separately in our method, as the output of our method is a probabilistic value, it is unnecessary to perform calibration by fitting a probability distribution as did in [30].”
MethodLearning discriminative trajectorylet detector sets for accurate skeleton-based action recognition
Qiao, R., Liu, L., Shen, C., & van den Hengel, A. (2017) · Pattern Recognition · doi:10.1016/j.patcog.2017.01.015 · Cited by 53
“Our approach firstly applies exemplar-SVM [10] to learn a large number of candidate detectors and then selects detectors according to their discriminative performance over the trajec-torylets in the training set.”
MethodOne-Shot Generative Domain Adaptation
Yang, C., Shen, Y., Zhang, Z., Xu, Y., Zhu, J., Wu, Z., & Zhou, B. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00711 · Cited by 65
“Regarding the one-shot target domain, the mechanism behind the classifier is very similar to Exemplar SVM (Malisiewicz et al., 2011), which also suggests that it is sufficient to obtain a good decision boundary with one positive and many negative samples.”
MethodCompRess: Self-Supervised Learning by Compressing RepresentationsPreprint
Koohpayegani, S. A., Tejankar, A., & Pirsiavash, H. (2020) · arXiv (Cornell University) · doi:10.13016/m2n1nd-9amp · Cited by 104
Synchronizing Object Detection: Applications, Advancements and Existing Challenges
Hosain, M. T., Zaman, A., Abir, M. R., Akter, S., Mursalin, S., & Khan, S. S. (2024) · IEEE Access · doi:10.1109/access.2024.3388889 · Cited by 70
“For a considerable amount of time, the HOG detector functioned as an essential component of many object detectors [34], [41] [42] and different computer vision applications.”
BackgroundLeveraging the Wisdom of the Crowd for Fine-Grained Recognition
Deng, J., Krause, J., Stark, M. M., & Fei-Fei, L. (2015) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2015.2439285 · Cited by 52
“All classifiers are based on HOG [40] features in connection with a one-versus-all linear SVM. Empirically, we found that exemplar SVMs [41] did not result in a significant improvement but were computationally much more demanding.”
BackgroundDynamic ensembles of exemplar-SVMs for still-to-video face recognition
Bashbaghi, S., Granger, É., Sabourin, R., & Bilodeau, G.-A. (2017) · Pattern Recognition · doi:10.1016/j.patcog.2017.04.014 · Cited by 52
“A method called exemplar-SVM (e-SVM) [49] has been proposed to train a separate SVM classifier with DEC for each individual of interest.”
MethodMultitask Learning for Object Localization With Deep Reinforcement Learning
Wang, Y., Zhang, L., Wang, L., & Wang, Z. (2018) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2018.2885813 · Cited by 54
“To localize object, the classic method is sliding windows approach [2]–[4].”
MethodLearning to Weight Samples for Dynamic Early-Exiting Networks
Han, Y., Pu, Y., Lai, Z., Wang, C., Song, S., Cao, J., Huang, W., Deng, C., & Huang, G. M. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20083-0_22 · Cited by 79
One-Shot Learning of Scene Locations via Feature Trajectory Transfer
Kwitt, R., Hegenbart, S., & Niethammer, M. (2016) · 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.2016.16 · Cited by 61
“Exemplar SVMs [16] are used to train detectors for the single gesture instances and the detectors are then used to mine more training data from the external database.”
MethodA Re-Balancing Strategy for Class-Imbalanced Classification Based on Instance Difficulty
Yu, S., Guo, J., Zhang, R., Fan, Y., Wang, Z., & Cheng, X. (2022) · 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/cvpr52688.2022.00017 · Cited by 68
“The most popular works quantify instances’ difficulty in terms of the losses incurred by the model [8, 10, 25, 29].”
BackgroundSuperpixel Graph Label Transfer with Learned Distance Metric
Gould, S. J., Zhao, J., He, X., & Zhang, Y. (2014) · Lecture notes in computer science · doi:10.1007/978-3-319-10590-1_41 · Cited by 58
“An extension of this work [5] includes cues derived from per-exemplar object detectors, which have been shown to work well for the object detection task [12].”
MethodAn Exemplar-Based Multi-View Domain Generalization Framework for Visual Recognition
Niu, L., Li, W., Xu, D., & Cai, J. (2016) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2016.2615469 · Cited by 50
“2) Baselines: We compare EMVDG_CO and EMVDG_MK methods with two basic baselines, i.e., SVM [51] and ESVM (ESVM) [22], as well as three sets of baseline methods: the multi-view learning approaches, the domain generalization approaches, and the latent domain discovering approaches.”
MethodInfluentialEyes in the Dark: Distributed Scene Understanding for Disaster Management
Li, L., Ota, K., Dong, M., & Borjigin, W. (2017) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2017.2740294 · Cited by 50
“The ODLN&Linear SVM [24] is 30.39%, the ODLN&DPano [18] is 37.16%, the SVM [39]&VTCN(with VI module) is 18.33%.”
MethodInfluentialRecognising Human-Object Interaction via Exemplar Based Modelling
Hu, J., Zheng, W., Lai, J., Gong, S., & Xiang, T. (2013) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2013.390 · Cited by 49
“Despite that exemplar based modelling has been applied to a variety of visual recognition problems including scene recognition [10], object detection [14], pose estimation [15], exemplar based HOI modelling has been mostly unexploited.”
MethodLSTM Self-Supervision for Detailed Behavior Analysis
Brattoli, B., Buchler, U., Wahl, A.-S., Schwab, M. E., & Ommer, B. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.399 · Cited by 49
“4 shows the average AUC using HOG-LDA[19], Exemplar-SVMs[29], Exemplar-CNN[13], CliqueCNN[8], and our approach.”
MethodA too-good-to-be-true prior to reduce shortcut reliance
Dagaev, N., Roads, B. D., Luo, X., Barry, D. N., Patil, K. R., & Love, B. C. (2022) · Pattern Recognition Letters · doi:10.1016/j.patrec.2022.12.010 · Cited by 61
“Some authors suggested to mitigate the impact of easy examples and focus on hard ones [27, 35].”
Small Object Detection via Precise Region-Based Fully Convolutional Networks
Zhang, D., Hu, J., Li, F., Ding, X., Sangaiah, A. K., & Sheng, V. S. (2021) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2021.017089 · Cited by 48
An Exemplar-Based CRF for Multi-instance Object Segmentation
He, X., & Gould, S. J. (2014) · 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.2014.45 · Cited by 50
“The exemplar SVMs are trained with reference images only and the reference masks are used to remove the background clutter.”
MethodInfluentialLearning a Unified Sample Weighting Network for Object Detection
Cai, Q., Pan, Y., Wang, Y., Liu, J., Yao, T., & Mei, T. (2020) · 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/cvpr42600.2020.01418 · Cited by 52
Deep Learning Methods for Abstract Visual Reasoning: A Survey on Raven's Progressive Matrices
Małkiński, M., & Mańdziuk, J. (2025) · ACM Computing Surveys · doi:10.1145/3715093 · Cited by 61
“Moreover, autoencoders were successfully utilised in [15] for learning transferable features in simple AVR tasks, whereas the authors of [21] incorporated an autoencoder as part of their architecture for creating so-called generative analogies.”
Sadgrove, E., Falzon, G., Miron, D., & Lamb, D. W. (2018) · Computers in Industry · doi:10.1016/j.compind.2018.03.014 · Cited by 47
Object-Aware Dense Semantic Correspondence
Yang, F., Li, X., Cheng, H., Li, J., & Chen, L. (2017) · 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.2017.442 · Cited by 50
“Since the solution only depends on a small set of “hard” negative support vectors [10], we use the hard-negative mining approach [20] to cope with millions of negative windows.”
MethodDeepCar 5.0: Vehicle Make and Model Recognition Under Challenging Conditions
Amirkhani, A., & Barshooi, A. H. (2022) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2022.3212921 · Cited by 52
Deep neural networks-based vehicle detection in satellite images
Jiang, Q., Cao, L., Cheng, M., Wang, C., & Li, J. (2015) · 2015 International Symposium on Bioelectronics and Bioinformatics (ISBB) · doi:10.1109/isbb.2015.7344954 · Cited by 45
“[6] propose a method named ExemplarSVMs(ESVM) for object detection.”
MethodAbstract Reasoning with Distracting Features
Zheng, K., Zha, Z.-J., & Wei, W. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.00569 · Cited by 56
“Curriculum learning The teaching strategies of weighting each training example have been well studied in the literature[5, 6, 18, 21, 28, 32].”
BackgroundA Novel Approach to Automatic Road-Accident Detection using Machine Vision Techniques
Ravindran, V., Viswanathan, L., & Rangaswamy, S. (2016) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2016.071130 · Cited by 44
“After extracting the HOG feature vectors, they are used to train an SVM model [15].”
MethodInfluentialMax-Margin Contrastive Learning
Shah, A., Sra, S., Chellappa, R., & Cherian, A. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i8.20796 · Cited by 62
Latent Hough Transform for Object Detection
Razavi, N., Gall, J., Kohli, P., & Gool, L. V. (2012) · Lecture notes in computer science · doi:10.1007/978-3-642-33712-3_23 · Cited by 43
“Instead of grouping the training examples, [26] train a model for every single positive instance in the training data.”
BackgroundModel recommendation: Generating object detectors from few samples
Wang, Y.-X., & Hebert, M. (2015) · 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.2015.7298770 · Cited by 43
“ESVMs are restricted in that each model is highly tuned for a specific instance, and thus constrained to one particular set of categories when building models.”
BackgroundMethodInfluentialFast object detection based on selective visual attention
Guo, M., Zhao, Y., Zhang, C., & Chen, Z. (2014) · Neurocomputing · doi:10.1016/j.neucom.2014.04.054 · Cited by 44
3D object detection: Learning 3D bounding boxes from scaled down 2D bounding boxes in RGB-D images
Rahman, M. M., Tan, Y., Xue, J., Shao, L., & Lü, K. (2018) · Information Sciences · doi:10.1016/j.ins.2018.09.040 · Cited by 44
Wu, X., Hong, D., Ghamisi, P., Li, W., & Tao, R. (2018) · Remote Sensing · doi:10.3390/rs10121990 · Cited by 42
“• The Exemplar-SVM detector [43] adopts template integration instead of a single template to realize object detection.”
MethodDelving into Sample Loss Curve to Embrace Noisy and Imbalanced Data
Jiang, S., Li, J., Wang, Y., Huang, B., Zhang, Z., & Xu, T. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i6.20661 · Cited by 53
Visualizing Object Detection Features
Vondrick, C., Khosla, A., Pirsiavash, H., Malisiewicz, T., & Torralba, A. B. (2016) · International Journal of Computer Vision · doi:10.1007/s11263-016-0884-7 · Cited by 42
“We optimize the above with SPAMS (Mairal et al, 2009).”
BackgroundInfluentialSelf-citationA deep survey on supervised learning based human detection and activity classification methods
Khan, M. A., Mittal, M., Goyal, L. M., & Roy, S. S. (2021) · Multimedia Tools and Applications · doi:10.1007/s11042-021-10811-5 · Cited by 45
“In [191] PASCAL VOC2007 used for multiple object detection.”
MethodSmall object detection in diverse application landscapes: a survey
Iqra, Giri, K. J., & Javed, M. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-18866-w · Cited by 48
“Inthe 2000s,machinelearningtechniqueswereintroduced,withanotablefocusonSupportVec-torMachines(SVMs)[22]and,subsequently,ensemblemethodslikeAdaBoost[23].”
Bird Part Localization Using Exemplar-Based Models with Enforced Pose and Subcategory Consistency
Liu, J., & Belhumeur, P. N. (2013) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2013.313 · Cited by 40
“A well-known detection method which had success on pedestrian [12] could not perform well on birds, as shown in [21].”
MethodContextual Exemplar Classifier-Based Image Representation for Classification
Zhang, C., Huang, Q., & Tian, Q. (2016) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2016.2527380 · Cited by 40
“[35] used an exemplar support vector machines for object detection in an ensemble way and then”
MethodLess Is Better: Unweighted Data Subsampling via Influence Function
Wang, Z., Zhu, H. C., Dong, Z., He, X., & Huang, S. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i04.6103 · Cited by 64
Hardness-Aware Deep Metric Learning
Zheng, W., Lu, J., & Zhou, J. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2020.2980231 · Cited by 51
“Hard negative mining has been employed in many machine learning tasks to enhance the training efficiency and boost performance, like supervised learning [13], [16], [35], [57], [58], exemplarbased learning [29] and unsupervised learning [1], [51].”
BackgroundGraph-Based Discriminative Learning for Location Recognition
Cao, S., & Snavely, N. (2014) · International Journal of Computer Vision · doi:10.1007/s11263-014-0774-9 · Cited by 40
“At the other extreme, one could learn a distance metric for each image in the database, analogous to how Exemplar SVMs have been used for object detection [20].”
MethodLabel Propagation from ImageNet to 3D Point Clouds
Wang, Y., Ji, R., & Chang, S.-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.2013.403 · Cited by 45
“Note that we do not leverage randomly sampled rectangles used in recent works in search based segmentation [14][21] or object detection [18], to ensure label consistency among pixels in each region, as widely assumed for superpixels [19]2.”
BackgroundMethodInfluentialComprehensive wheat lodging detection after initial lodging using UAV RGB images
Azizi, A., Zhang, Z., Rui, Z., Li, Y., Igathinathane, C., Flores, P., Mathew, J. J., Pourreza, A., Han, X., & Zhang, M. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.121788 · Cited by 39
Expanded Parts Model for Semantic Description of Humans in Still Images
Sharma, G., Jurie, F., & Schmid, C. (2016) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2537325 · Cited by 41
“Although SPM has never been viewed as a template learning method, methods using gradients based features [25], [26], [27], [28] have”
Fused One-vs-All Features With Semantic Alignments for Fine-Grained Visual Categorization
Zhang, X., Xiong, H., Zhou, W., & Tian, Q. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2509425 · Cited by 41
“Recently, Christoph et al. [39] suggested to transfer part annotations from objects via performing a simple but very powerful global matching and a subsequent ensemble learning.”
BackgroundImproving CNN linear layers with power mean non-linearity
Zhang, C.-L., & Wu, J. (2018) · Pattern Recognition · doi:10.1016/j.patcog.2018.12.029 · Cited by 40
Chen, L., Li, G., Xie, W., Tan, J., Li, Y., Pu, J., Chen, L., Gan, D., & Shi, W. (2024) · Energies · doi:10.3390/en17205177 · Cited by 47
Fernando, B., & Tuytelaars, T. (2013) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2013.316 · Cited by 38
“It does not require any negative images – as opposed to exemplar SVM [2, 14], Joint-SVM [2] or the method presented in [12].”
MethodInfluentialStructured Weak Semantic Space Construction for Visual Categorization
Zhang, C., Cheng, J., & Tian, Q. (2017) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2017.2728060 · Cited by 38
“[30] combined exemplar support vector machine (SVM) classifiers for object detection while Zhang et al.”
MethodExemplar-based facial expression recognition
Farajzadeh, N., & Hashemzadeh, M. (2018) · Information Sciences · doi:10.1016/j.ins.2018.05.057 · Cited by 38
Crop-saving with AI: latest trends in deep learning techniques for plant pathology
Salman, Z., Muhammad, A., Piran, M. J., & Han, D. (2023) · Frontiers in Plant Science · doi:10.3389/fpls.2023.1224709 · Cited by 40
“Islam M (Islam et al., 2017). has used image segmentation with multiclass SVM on RGB based features of 300 images acquired from Plant Village dataset and achieved an accuracy of 95% to classify 2 diseases; late blight and early blight in potato leaves.”
MethodInfluentialEnhancing Exemplar SVMs using Part Level Transfer Regularization
Aytar, Y., & Zisserman, A. (2012) · British Machine Vision Conference (BMVC) · doi:10.5244/c.26.79 · Cited by 37
“Transfer learning [13, 28], has been applied to computer vision primarily for image classification [17, 23, 24, 28], rather than detection, and we discuss the relation of the EE-SVM formulation to the standard objectives of transfer learning in section 2.”
BackgroundMethodInfluentialView-based 3D object retrieval with discriminative views
Wang, D., Wang, B., Zhao, S., Yao, H., & Liu, H. (2017) · Neurocomputing · doi:10.1016/j.neucom.2016.06.095 · Cited by 37
Qiu, S., Wen, G., & Fan, Y. (2017) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2017.2655098 · Cited by 38
Querying for video events by semantic signatures from few examples
Mazloom, M., Habibian, A., & Snoek, C. G. M. (2013) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/2502081.2502160 · Cited by 41
“We also report results using an exemplar-SVM [7] trained on each individual positive query and all available negative examples, and then”
BackgroundMethodRenewing the respect for similarity
Edelman, S., & Shahbazi, R. (2012) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2012.00045 · Cited by 38
“This calls for alternatives such as Exemplar-SVM (Malisiewicz et al., 2011), which may, perhaps, be amenable to implementation by augmenting RBF units with massive inhibition (Wang et al.”
MethodVisual object tracking by correlation filters and online learning
Zhang, X., Xia, G.-S., Lu, Q., Shen, W., & Zhang, L. (2017) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2017.07.009 · Cited by 35
Mapping Brick Kilns to Support Environmental Impact Studies around Delhi Using Sentinel-2
Misra, P., Imasu, R., Hayashida, S., Arbain, A. A., Avtar, R., & Takeuchi, W. (2020) · ISPRS International Journal of Geo-Information · doi:10.3390/ijgi9090544 · Cited by 35
“Regions of brick kiln locations overlapping more than 70% were removed during post-processing through ‘faster non-max suppression’ algorithm [51].”
MethodBi-directional Contrastive Learning for Domain Adaptive Semantic Segmentation
Lee, G., Eom, C., Lee, W.-K., Park, H.-K., & Ham, B. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20056-4_3 · Cited by 43
Dynamic belief fusion for object detection
Lee, H., Kwon, H., Robinson, R. M., Nothwang, W. D., & Marathe, A. R. (2016) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2016.7477574 · Cited by 39
“The proposed DBF algorithm was evaluated on the ARL dataset and its average precision (AP) was compared to that of four individual detection algorithms (the HOG-SVM detector was not utilized on the ARL dataset.) and four other “baseline” fusion methods (Platt, Bayes, WS, and DST) for each object class.”
BackgroundMethodInfluentialTraffic flow from a low frame rate city camera
Toropov, E., Gui, L.-Y., Zhang, S., Kottur, S., & Moura, J. M. F. (2015) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2015.7351516 · Cited by 34
“We threshold the candidates based on expected size, expected ratio, and proximity to image boundary, and use standard non-maximum suppression [18] to generate a final set of detections.”
MethodZagoris, K., Pratikakis, I., & Gatos, B. (2017) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2700721 · Cited by 34
“It is worth to note that the EWS [42] method uses the exemplar-SVM framework [44], and relies on a sliding-window search to retrieve the document regions that are likely to contain the query word.”
MethodLiu, W., Yu, Z., Lu, L., Wen, Y., Li, H., & Zou, Y. (2015) · Pattern Recognition · doi:10.1016/j.patcog.2015.04.014 · Cited by 33
“For example, if the query is located near decision boundary, then these local atoms play the role similar to support vectors in an SVM, or in an extreme case, exemplars in an exemplar SVM [17].”
BackgroundEnriching object detection with 2D-3D registration and continuous viewpoint estimation
Choy, C., Stark, M. M., Corbett‐Davies, S., & Savarese, S. (2015) · 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.2015.7298866 · Cited by 33
“First, it can be run in isolation, as a sliding window object detector similar in spirit to the exemplar SVM [16].”
BackgroundAdding Unlabeled Samples to Categories by Learned Attributes
Choi, J., Rastegari, M., Farhadi, A., & Davis, L. S. (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.2013.118 · Cited by 39
“For our purposes, though, we can accomplish the same thing by analyzing how the ranks of unlabeled samples change when a single sample is eliminated from the training set of the attribute SVM.”
BackgroundInfluentialDeep unsupervised learning of visual similarities
Sanakoyeu, A., Bautista, M. Á., & Ommer, B. (2018) · Pattern Recognition · doi:10.1016/j.patcog.2018.01.036 · Cited by 37
“HOG-LDA [6] Ex-SVM [1] Ex-CNN [15] Alexnet [”
InfluentialVisual tracking via exemplar regression model
Ma, X., Liu, Q., He, Z., Zhang, X., & Chen, W.-S. (2016) · Knowledge-Based Systems · doi:10.1016/j.knosys.2016.05.028 · Cited by 32
“We borrow the concept of exemplar from Exemplar-SVM [29].”
MethodInfluentialTraining Deformable Part Models with Decorrelated Features
Girshick, R., & Malik, J. (2013) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2013.375 · Cited by 34
“As a reference point, exemplar SVM [17] does only slightly better, 19.”
BackgroundMethodInfluentialExemplar Driven Character Recognition in the Wild
Sheshadri, K., & Divvala, S. (2012) · British Machine Vision Conference (BMVC) · doi:10.5244/c.26.13 · Cited by 32
“In the spirit of [10], we adopt an “on the fly” calibration method, using positives selected by each exemplar based on SVM scores.”
BackgroundMethodInfluentialVisual Object Recognition with 3D-Aware Features in KITTI Urban Scenes
Yebes, J. J., Bergasa, L. M., & García-Garrido, M. Á. (2015) · Sensors · doi:10.3390/s150409228 · Cited by 36
“In contrast, [25] announced Exemplar-SVMs, which are compositional models trained from only one positive example and millions of negative ones.”
BackgroundVehicle Perception From Satellite
Zhao, B., Han, P., & Li, X. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3335953 · Cited by 31
“Traditional approaches follow a pipeline of region selection ( e.g. , superpixels [18], sliding window [19], selective search [20]), feature extraction ( e.g. , SIFT [21], HOG [22]) and classifier ( e.g. , SVM [23], Adaboost [24]).”
MethodLearning multi-level features for sensor-based human action recognition
Xu, Y. L., Shen, Z., Zhang, X., Gao, Y., Deng, S., Wang, Y., Fan, Y., & Chang, E. (2017) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2017.07.001 · Cited by 32
“For those reasons, generalized features that are directly learned from action perspectives would compromise to the common characteristic shared by different action patterns, resulting in ambiguous feature representations [30].”
BackgroundGuo, X., Liu, D., Jou, B., Zhu, M., Cai, A., & Chang, S. (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.2013.412 · Cited by 33
“• Ensemble of Exemplar-SVMs (ESVMs) [18].”
BackgroundMethodInfluentialExemplar-Based Recognition of Human–Object Interactions
Hu, J., Zheng, W., Lai, J., Gong, S., & Xiang, T. (2015) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2015.2397200 · Cited by 37
“recognition problems including scene recognition [36], object detection [37], and pose estimation [38], few works exploit exemplar-based modeling for HOI recognition.”
BackgroundSentiment analysis on social campaign “Swachh Bharat Abhiyan” using unigram method
Tayal, D. K., & Yadav, S. K. (2016) · AI & Society · doi:10.1007/s00146-016-0672-5 · Cited by 33
“Besides the advantages of SVM, the main drawback of using it is its high algorithmic complexity and extensive memory requirements (Malisiewicz et al. 2011).”
BackgroundImage-Specific Classification With Local and Global Discriminations
Zhang, C., Cheng, J., Li, C., & Tian, Q. (2017) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2017.2748952 · Cited by 31
“Instead of learning class-level classifiers, the use of exemplar classifiers [16]–[19] also became popular.”
Rangesh, A., & Trivedi, M. M. (2018) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2018.2804170 · Cited by 33
CollageParsing: Nonparametric Scene Parsing by Adaptive Overlapping Windows
Tung, F., & Little, J. J. (2014) · Lecture notes in computer science · doi:10.1007/978-3-319-10599-4_33 · Cited by 33
“Tighe and Lazebnik [33] recently extended the SuperParsing algorithm with per-exemplar detectors (Exemplar-SVMs [23]).”
BackgroundGenerative negative replay for continual learning
Graffieti, G., Maltoni, D., Pellegrini, L., & Lomonaco, V. (2023) · Neural Networks · doi:10.1016/j.neunet.2023.03.006 · Cited by 34
Towards 3D object detection with bimodal deep Boltzmann machines over RGBD imagery
Liu, W., Ji, R., & Li, S. (2015) · 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.2015.7298920 · Cited by 31
“The process of detecting objects of interest is summarized in Algorithm 1.”
MethodExemplar SVMs as visual feature encoders
Zepeda, J., & Pérez, P. (2015) · 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.2015.7298924 · Cited by 30
“[20] proposed using E-SVMs to transfer meta-data (e.”
MethodDeep Learning Based Semantic Video Indexing and Retrieval
Podlesnaya, A., & Podlesnyy, S. (2017) · Lecture notes in networks and systems · doi:10.1007/978-3-319-56991-8_27 · Cited by 34
Deep Unsupervised Similarity Learning Using Partially Ordered Sets
Bautista, M. Á., Sanakoyeu, A., & Ommer, B. (2017) · 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.2017.208 · Cited by 32
“In order to do so, we take a sequence from HOG-LDA [12] Ex-SVM [16] Ex-CNN [6] 0.62 0.72 0.64 Alexnet [14] Doersch et. al [5] Suffle&Learn [20] 0.65 0.62 0.63 CliqueCNN [3] Ours scratch Ours Imagenet 0.83 0.78 0.85 Table 1.”
InfluentialLearning Object Detectors With Semi-Annotated Weak Labels
Zhang, D., Han, J., Guo, G., & Zhao, L. (2018) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2018.2884173 · Cited by 31
“AS A FUNDAMENTAL problem in the computer vision community, object detection [1], [2] has received wide attention in the last few decades.”
BackgroundDBF: Dynamic Belief Fusion for Combining Multiple Object Detectors
Lee, H., & Kwon, H. (2019) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2952847 · Cited by 31
“By equally treating prior probability of target and non-target (i.e., mðT Þ ¼ mð:T Þ ¼ 12), we consider the fusion score as below: mfðcÞ ¼ mðT jcÞ mð:T jcÞ ¼ 1 2 YK k¼1 lkðckjT Þ 1 2 YK k¼1 lkðckj:T Þ: (15) TABLE 1 Average Precision on the ARL Dataset [14] mAP AP per category chair container door poster stair DSIFT .0…”
BackgroundMethodInfluentialVision-Based Human Detection Techniques: A Descriptive Review
Sumit, S. S., Rambli, D. R. A., & Mirjalili, S. (2021) · IEEE Access · doi:10.1109/access.2021.3063028 · Cited by 35
“ChallengesThe algorithm performedwell in case of human but could not perform well on detecting objects with high articulation such as birds [74].”
BackgroundA Joint Deep Boltzmann Machine (jDBM) Model for Person Identification Using Mobile Phone Data
Alam, M. R., Bennamoun, M., Togneri, R. B., & Sohel, F. A. (2016) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2016.2615524 · Cited by 28
“In [16], a cross-modality DBM framework was used along with an Exemplar-SVM [17] for 3D scenes object detection.”
MethodFusion of an Ensemble of Augmented Image Detectors for Robust Object Detection
Wei, P., Ball, J. E., & Anderson, D. T. (2018) · Sensors · doi:10.3390/s18030894 · Cited by 33
“[31] trained a linear SVM for each class in the training set.”
MethodMulti-Feature View-Based Shallow Convolutional Neural Network for Road Segmentation
Junaid, M., Ghafoor, M., Hassan, A., Khalid, S., Tariq, S. A., Ahmed, G., & Zia, T. (2020) · IEEE Access · doi:10.1109/access.2020.2968965 · Cited by 28
“Similarly, ML techniques have been also used in object detection [25], [26], classification [27], [28], and segmentation [29]–[31].”
MethodMulti-spectral template matching based object detection in a few-shot learning manner
Chen, F., Cao, Z., Xiao, Y., Fang, Z., & Zhou, J. T. (2022) · Information Sciences · doi:10.1016/j.ins.2022.12.067 · Cited by 28
Location recognition over large time lags
Fernando, B., Tommasi, T., & Tuytelaars, T. (2015) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2015.05.016 · Cited by 33
“way indicate this approach is clearly not suitable for the task at hand. We consider as reference also the method proposed in [36] based on the combination of HOG features [8] and Exemplar SVM (ESVM, [26]). Its performance is quite low: as evident in the examples shown in Figure 2, the HOG features mostly focus on the…”
MethodRobust vehicle detection by combining deep features with exemplar classification
Cao, L., Jiang, Q., Cheng, M., & Wang, C. (2016) · Neurocomputing · doi:10.1016/j.neucom.2016.03.094 · Cited by 28
“[12] proposed to use Exemplar-SVMs (ESVM) for object detection.”
MethodCKD: Contrastive Knowledge Distillation From a Sample-Wise Perspective
Zhu, W., Zhou, X., Zhu, P.-F., Wang, Y., & Hu, Q.-H. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2025.3573474 · Cited by 29
“Some works have revealed that each image is distinctive and can vary significantly from other images within the same semantic class [48], [80].”
Significance and Limitations of Deep Neural Networks for Image Classification and Object Detection
Sood, S., Singh, H., Malarvel, M., & Ahuja, R. (2021) · International Conference on Smart Electronics and Communication (ICOSEC) · doi:10.1109/icosec51865.2021.9591759 · Cited by 27
Adaptive region pooling for object detection
Tsai, Y., Hamsici, O. C., & Yang, M. (2015) · 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.2015.7298673 · Cited by 27
“# of models # of windows per model # of proposals ESVM [29] all exemplars sliding windows > 106 LDA [20] < 100 sliding windows > 106 Our method 50 20 103 rithm with other exemplar-based methods [29, 20].”
BackgroundMethodVisual Phrases for Exemplar Face Detection
Kumar, V., Namboodiri, A., & Jawahar, C. V. (2015) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.231 · Cited by 26
Object categorization in sub-semantic space
Zhang, C., Cheng, J., Liu, J., Pang, J., Liang, C., Huang, Q., & Tian, Q. (2014) · Neurocomputing · doi:10.1016/j.neucom.2014.03.059 · Cited by 26
Online Domain Adaptation for Multi-Object Tracking
Gaidon, A., & Vig, E. (2015) · British Machine Vision Conference (BMVC) · doi:10.5244/c.29.3 · Cited by 28
“Indeed, once the detectors w i are updated in frame t , a new scene-adapted detector is readily available as: As we use linear classifiers, this multi-task learning is akin to a “fusion” of exemplar-based models ( e.g., Exemplar-SVMs [30]).”
MethodMetaAugment: Sample-Aware Data Augmentation Policy Learning
Zhou, F., Li, J., Xie, C., Chen, F., Hong, L., Sun, R., & Li, Z. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i12.17324 · Cited by 35
Metadata
- Authors
- Malisiewicz, T., Gupta, A., & Efros, A. A.
- Affiliations
- Carnegie Mellon University (US)
- Venue
- 2011 International Conference on Computer Vision
- Publisher
- IEEE
- Published
- 2011-11
- Pages
- 89-96
- OpenAlex
- W1989684337
- Cited by
- 865
- References
- 25
- Citations by year
- 2026: 4 · 2025: 18 · 2024: 16 · 2023: 29 · 2022: 54 · 2021: 52 · 2020: 49 · 2019: 45 · 2018: 65 · 2017: 78 · 2016: 105 · 2015: 148 · 2014: 97 · 2013: 65 · 2012: 36
- FWCI
- 48.23
- Type
- conference paper
- Language
- en
- Open access
- green
- figshare.com
- Licence
- cc-by
- Locations
- IEEE/CVF International Conference on Computer Vision (ICCV) (publishedVersion)
- Figshare (submittedVersion, cc-by)
- KiltHub Repository (submittedVersion)
- KiltHub Repository
- Funders
- National Science Foundation (0905402)
- Multidisciplinary University Research Initiative
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Image and Video Retrieval Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Visual Attention and Saliency Detection
- Keywords
- object detection, linear SVM classifier, Pascal VOC, scene understanding, part-based models, generalization
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.