High-dimensional signature compression for large-scale image classification
The DOI matches its Crossref record; 33 of the 34 references it cites were found in a registry or an index.
Abstract
We address image classification on a large-scale, i.e. when a large number of images and classes are involved. First, we study classification accuracy as a function of the image signature dimensionality and the training set size. We show experimentally that the larger the training set, the higher the impact of the dimensionality on the accuracy. In other words, high-dimensional signatures are important to obtain state-of-the-art results on large datasets. Second, we tackle the problem of data compression on very large signatures (on the order of 105dimensions) using two lossy compression strategies: a dimensionality reduction technique known as the hash kernel and an encoding technique based on product quantizers. We explain how the gain in storage can be traded against a loss in accuracy and/or an increase in CPU cost. We report results on two large databases - ImageNet and a dataset of lM Flickr images - showing that we can reduce the storage of our signatures by a factor 64 to 128 with little loss in accuracy. Integrating the decompression in the classifier learning yields an efficient and scalable training algorithm. On ILSVRC2010 we report a 74.3% accuracy at top-5, which corresponds to a 2.5% absolute improvement with respect to the state-of-the-art. On a subset of 10K classes of ImageNet we report a top-1 accuracy of 16.7%, a relative improvement of 160% with respect to the state-of-the-art.
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
“since then the best published results are 45.7% and 25.7% with an approach that averages the predictions of two classifiers trained on Fisher Vectors (FVs) computed from two types of densely sampled features.29”
From ImageNet classification with deep convolutional neural networks (2017) · cites this paper as [29]
Citations
Cites
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 1×
“org) consists of more than 12M images of 17K concepts [7] and Flickr contains thousands of groups (www.”
BackgroundBeyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories
Lazebnik, S., Schmid, C., & Ponce, J. (2006) · 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Volume 2 (CVPR'06) · doi:10.1109/cvpr.2006.68 · Cited by 7,916 · cited 1×
“[15], one can split an image into several regions, compute one FV per region and concatenate the per-region FVs.”
BackgroundLocality-constrained Linear Coding for image classification
Wang, J., Yang, J., Yu, K., Lv, F., Huang, T., & Gong, Y. (2010) · 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2010.5540018 · Cited by 3,100 · cited 2×
“Maji and Berg [18] and Wang et al. [31] then proposed efficient algorithms to learn IKSVMs.”
ResultProduct Quantization for Nearest Neighbor Search
Jégou, H., Douze, M., & Schmid, C. (2011) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2010.57 · Cited by 2,956 · cited 3×
“[13] proposed product quantizers (PQ) for NN-search of SIFT vectors.”
BackgroundMethodLinear spatial pyramid matching using sparse coding for image classification
Yang, J., Yu, K., Gong, Y., & Huang, T. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206757 · Cited by 2,828 · cited 1×
“Yang et al. [36], Wang et al. [32] and Boureau et al. [4] showed that replacing the average pooling stage in the BOV computation by a max-pooling yielded excellent results.”
ResultWeiss, Y., Torralba, A. B., & Fergus, R. (2008) · neural information processing systems · Cited by 2,767
Aggregating local descriptors into a compact image representation
Jegou, H., Douze, M., Schmid, C., & Perez, P. (2010) · 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2010.5540039 · Cited by 2,692 · cited 3×
“For instance, [23] requires dense projections and [34, 5, 14] require a global PCA.”
BackgroundMethodImproving the Fisher Kernel for Large-Scale Image Classification
Perronnin, F., Sánchez, J., & Mensink, T. (2010) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-15561-1_11 · Cited by 2,588 · cited 15×
“As shown in [22], square-rooting and L2-normalizing the FV can greatly enhance the classification accuracy.”
BackgroundMethodInfluentialSelf-citationZhang, J., Marszałek, M., Lazebnik, S., & Schmid, C. (2007) · International Journal of Computer Vision · doi:10.1007/s11263-006-9794-4 · Cited by 1,923 · cited 1×
“The winners of the PASCAL VOC 2007 [8] and 2008 [9] competitions used a similar paradigm: many types of low-level local features are extracted (referred to as “channels”), one bag-of-visual-words (BOV) histogram is computed for each channel and non-linear kernel classifiers such as SVMs are used to perform classificat…”
MethodEvaluating Color Descriptors for Object and Scene Recognition
van de Sande, K. E. A., Gevers, T., & Snoek, C. G. M. (2010) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2009.154 · Cited by 1,783 · cited 1×
“The winners of the PASCAL VOC 2007 [8] and 2008 [9] competitions used a similar paradigm: many types of low-level local features are extracted (referred to as “channels”), one bag-of-visual-words (BOV) histogram is computed for each channel and non-linear kernel classifiers such as SVMs are used to perform classificat…”
Method
Research works printed but not linked: 6
- shi, 2009, Hash kernels, AISTATS
- bottou, 0, SGD
- berg, 2010, ILSVRC
- gong, 2009, Image classification using Gaussian mixture and local coordinate coding, PASCAL VOC 2009 Workshop
- everingham, 0, The PASCAL VOC 2008 Results
- everingham, 2007, The PASCAL VOC
Cited by
Items 101–200 of 272
Zhang, Y., Wu, J., & Cai, J. (2016) · IEEE Transactions on Image Processing · doi:10.1109/tip.2016.2549360 · Cited by 28
Deep Network Classification by Scattering and Homotopy Dictionary LearningPreprint
Zarka, J., Thiry, L., Angles, T., & Mallat, S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1910.03561 · Cited by 49
“Prior to convolutional networks, SIFT representations combined with Fisher vector encoding reached a Top 5 classification accuracy of 74.3% with multiple model averaging (Sánchez & Perronnin, 2011).”
MethodInfluentialAmbikapathy, A., & Singh, A. V. (2017) · Lecture notes in electrical engineering · doi:10.1007/978-981-10-4762-6_27 · Cited by 26
Fried Binary Embedding for High-Dimensional Visual Features
Hong, W., Yuan, J., & Bhattacharjee, S. D. (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.2017.659 · Cited by 25
“For the first challenge, it has been noticed that for input feature vector of dimensionality d, the length b of binary codes required to achieve reasonable accuracy is usually O(d) [17, 3, 25].”
BackgroundGe, T., He, K., & Sun, J. Z. (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.2014.125 · Cited by 25
“The product quantization techniques have witnessed great success in large scale problems, including nearest neighbor search [14, 3, 9, 10, 22, 30] and large scale learning [25, 27].”
BackgroundRegion Ranking SVM for Image Classification
Wei, Z., & Hoai, 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.2016.326 · Cited by 27
Yoo, C.-H., Kim, S., Jung, J.-Y., & Ko, S.-J. (2017) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2017.02.009 · Cited by 24
Projection Bank: From High-Dimensional Data to Medium-Length Binary Codes
Liu, L., Yu, M., & Shao, L. (2015) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.323 · Cited by 25
“For Flickr 1M, we downloaded close to one million web images with 55 groups from Flickr inspired by [30, 27].”
BackgroundMethodInfluentialLand-use scene classification: a comparative study on bag of visual word framework
Shahriari, M., & Bergevin, R. (2017) · Multimedia Tools and Applications · doi:10.1007/s11042-016-4316-z · Cited by 25
“Dense features have shown to be an essential component of many state-of-the-art classification approaches [16, 26, 31].”
BackgroundMethodSong, W., Wang, Y., & Pan, Z. (2023) · Information Sciences · doi:10.1016/j.ins.2023.119216 · Cited by 22
Do, T. (2014) · Vietnam Journal of Computer Science · doi:10.1007/s40595-013-0013-2 · Cited by 25
“The recent work [20] and [10] study the impact of high-dimensional Fisher vectors on large-datasets.”
BackgroundA-DeepPixBis: Attentional Angular Margin for Face Anti-Spoofing
Hossain, M. S., Rupty, L. K., Roy, K., Hasan, M., Sengupta, S., & Mohammed, N. (2020) · International Conference on Digital Image Computing: Techniques and Applications (DICTA) · doi:10.1109/dicta51227.2020.9363382 · Cited by 29
“Boulkenafet et al [4] performs Fisher Vector [13], [14], [15] encoding on SURF (Speeded-up Robust Features) on different color spaces.”
MethodVRFP: On-the-Fly Video Retrieval Using Web Images and Fast Fisher Vector Products
Han, X., Singh, B., Morariu, V. I., & Davis, L. S. (2017) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2017.2671414 · Cited by 25
“FVs constructed from SIFT features extracted on images are dense (around 50% of the entries are non-zero) [33] due to significant variation in appearance of small patches within an image.”
BackgroundHigh-Resolution Class Activation Mapping
Tagaris, T., Sdraka, M., & Stafylopatis, A. (2019) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2019.8803474 · Cited by 20
“1% top-5 error rate on ILSVRC 2014, a score which would be considered unsatisfactory even during the pre-deep learning era [13].”
BackgroundRandom projection-based partial feature extraction for robust face recognition
Ma, C., Jung, J.-Y., Kim, S.-W., & Ko, S.-J. (2014) · Neurocomputing · doi:10.1016/j.neucom.2014.09.004 · Cited by 19
CTA: Hardware-Software Co-design for Compressed Token Attention Mechanism
Wang, H., Xu, H., Wang, Y., & Han, Y. (2023) · IEEE International Symposium on High-Performance Computer Architecture (HPCA) · doi:10.1109/hpca56546.2023.10070997 · Cited by 25
Learning to Hash for Efficient Search Over Incomplete Knowledge Graphs
Wang, M., Shen, H., Wang, S., Yao, L., Jiang, Y., Qi, G., & Chen, Y. (2019) · Proceedings · doi:10.1109/icdm.2019.00174 · Cited by 16
“Most of learning-to-hash methods focused on similarity preserving and large scale image search/classification [10], [11], but cannot be simply applied to KGs which have important structure information.”
BackgroundByeon, W., Liwicki, M., & Breuel, T. M. (2015) · Pattern Recognition Letters · doi:10.1016/j.patrec.2015.06.003 · Cited by 16
“Following a standard dense-SIFT and PHOW feature (a variant of dense-SIFT descriptors, extracted at multiple scales) [5], multiple encoding schemes, including bag-of-visual-words (BoVW) and Fisher Vector (FV) [31,34], were compared.”
MethodRGB-D Image Multi-Target Detection Method Based on 3D DSF R-CNN
Hu, Q., & Zhai, L. (2018) · International Journal of Pattern Recognition and Artificial Intelligence · doi:10.1142/s0218001419540260 · Cited by 17
Efficient On-the-fly Category Retrieval Using ConvNets and GPUs
Chatfield, K., Simonyan, K., & Zisserman, A. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-16865-4_9 · Cited by 15
“Product quantization (PQ) has been widely used as a compression method for image features [15], [27], and works by splitting the original feature into Q-dimensional sub-blocks, each of which is encoded using a separate vocabulary of cluster centres pre-learned from a training set.”
Using line segments to train multi-stream stacked autoencoders for image classification
Tang, X., Hao, K., Wei, H., & Ding, Y. (2017) · Pattern Recognition Letters · doi:10.1016/j.patrec.2017.05.025 · Cited by 14
Movement Tube Detection Network Integrating 3D CNN and Object Detection Framework to Detect Fall
Zou, S., Min, W., Liu, L., Wang, Q., & Zhou, X. (2021) · Electronics · doi:10.3390/electronics10080898 · Cited by 17
“2 percent lower than that of the method where linear SVMs are trained on Fisher Vectors (FVs) computed from two types of densely sampled features [27].”
MethodMixture factor analysis with distance metric constraint for dimensionality reduction
Chen, J., Liao, L., Zhang, W., & Du, L. (2021) · Pattern Recognition · doi:10.1016/j.patcog.2021.108156 · Cited by 14
Optimization of Robust Loss Functions for Weakly-Labeled Image Taxonomies
McAuley, J., Arnau Ramisa, A., & Caetano, T. S. (2012) · International Journal of Computer Vision · doi:10.1007/s11263-012-0561-4 · Cited by 17
“Perhaps it is for this reason that many of the state-of-the-art methods in the ImageNet Challenge consisted of binary classifiers, such as multiclass SVMs, that merely optimized the score of a single prediction (Lin et al. 2011; Sánchez and Perronnin 2011).”
InfluentialDocument Image Classification, with a Specific View on Applications of Patent Images
Csurka, G. (2017) · The Kluwer international series on information retrieval · doi:10.1007/978-3-662-53817-3_12 · Cited by 14
Bilinear Optimized Product Quantization for Scalable Visual Content Analysis
Yu, L., Huang, Z. H., Shen, F., Song, J., Shen, H. T., & Zhou, X. (2017) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2722224 · Cited by 13
A distributed approach for large-scale classifier training and image classification
Mei, K., Dong, P., Hao, L., & Jianping Fan, B. (2014) · Neurocomputing · doi:10.1016/j.neucom.2014.04.042 · Cited by 12
LUWA Dataset: Learning Lithic Use-Wear Analysis on Microscopic Images
Zhang, J., Fang, I., Wu, H., Kaushik, A., Rodriguez, A., Zhao, H., Zhang, J., Zheng, Z., Йовита, Р., & Feng, C. (2024) · 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/cvpr52733.2024.02129 · Cited by 21
“We deploy classic methods such as SIFT+FVs [49], ResNets[24] and cutting-edge models such as ViT [17], ConvNeXts [34] and DINOv2 [13, 43] to our LUWA dataset.”
MethodExtremely Low Bit-Rate Nearest Neighbor Search Using a Set Compression Tree
Arandjelović, R., & Zisserman, A. (2014) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2014.2339821 · Cited by 15
“It was originally applied to SIFT vector matching, but has since been employed in large scale category retrieval [5] and used with inverted indexes for immediate retrieval [6], [7] (again addressing problem (ii)).”
MethodLarge scale multi-class classification with truncated nuclear norm regularization
Hu, Y., Jin, Z., Shi, Y., Zhang, D., Cai, D., & He, X. (2014) · Neurocomputing · doi:10.1016/j.neucom.2014.06.073 · Cited by 13
Tensorized Projection for High-Dimensional Binary Embedding
Hong, W., Meng, J., & Yuan, J. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.11292 · Cited by 13
“The results also validate that longer codes in general achieve better performance than shorter ones (Sánchez and Perronnin 2011).”
BackgroundResultInfluentialFisher Vectors for Fine-Grained Visual Categorization
Sánchez, J., Perronnin, F., & Akata, Z. (2011) · HAL · Cited by 12
Self-citationImproved Statistically Based Retrievals via Spatial-Spectral Data Compression for IASI Data
García-Sobrino, J., Laparra, V., Serra-Sagrista, J. S., Calbet, X., & Camps‐Valls, G. (2019) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2019.2901396 · Cited by 12
“Sánchez and Perronnin evaluated two lossy compression techniques, one based on product quantizers and another based on dimensionality reduction, to assess large-scale image classification in [5].”
BackgroundRelaxed Asymmetric Deep Hashing Learning: Point-to-Angle Matching
Li, J., Zhang, B., Lu, G., You, J., Xu, Y. D., Wu, F., & Zhang, D. D. (2019) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2019.2958061 · Cited by 13
“, image retrieval [1]–[3], pattern recognition [4]–[6], and data fusion [7].”
BackgroundTrace Augmentation: What Can Be Done Even Before Preprocessing in a Profiled SCA?
Pu, S., Yu, Y., Wang, W., Guo, Z., Liu, J., Gu, D., Wang, L., & Gan, J. (2018) · Lecture notes in computer science · doi:10.1007/978-3-319-75208-2_14 · Cited by 12
PixNet: A Localized Feature Representation for Classification and Visual Search
Pourian, N., & Manjunath, B. S. (2015) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2015.2410734 · Cited by 12
Large scale classifiers for visual classification tasks
Doan, T.-N., Do, T., & Poulet, F. (2014) · Multimedia Tools and Applications · doi:10.1007/s11042-014-2049-4 · Cited by 13
“A number of prominent works can be found in [7, 28, 36, 40] where the data is first explicitly transformed from linear space to non-linear space to ensure linear separability of the classes and then linear classifier is trained in the resulting space—non-linear space (usually of a much higher dimension).”
BackgroundChart classification by combining deep convolutional networks and deep belief networks
Liu, X., Tang, B., Wang, Z., Xu, X., Pu, S., Tao, D., & Song, M. (2015) · IAPR International Conference on Document Analysis and Recognition (ICDAR) · doi:10.1109/icdar.2015.7333872 · Cited by 11
“The Fisher Vectors (FVs)[16] are then used as the classifier.”
MethodCombining multi-representation for multimedia event detection using co-training
Bin, Y., Yang, Y., Shen, F., & Xu, X. (2016) · Neurocomputing · doi:10.1016/j.neucom.2016.03.091 · Cited by 11
Recognition of radio signals with deep learning Neural Networks
Işık, G., & Artuner, H. (2016) · Signal Processing and Communications Applications Conference (SIU) · doi:10.1109/siu.2016.7495870 · Cited by 13
“Eğitim boyunca CNN, filtrelerin içeriğini (öznitelikleri), gerçekleştirmesi istenen işleme göre kendisi otomatik olarak öğrenir.”
InfluentialUnderstanding and localizing activities from correspondences of clustered trajectories
Turchini, F., Seidenari, L., & Bimbo, A. D. (2016) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2016.11.007 · Cited by 12
Mapping RNA-seq reads to transcriptomes efficiently based on learning to hash method
Yu, X., & Liu, X. (2019) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2019.103539 · Cited by 12
Enabling low bitrate mobile visual recognition
Su, Y.-C., Chiu, T.-H., Chen, Y., Yeh, C.-Y., & Hsu, W. H. (2013) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/2502081.2502110 · Cited by 9
“We also examined the optimal scales up for thumbnail images on feature extraction, and the result is in favor of scaling up image to the original size.”
BackgroundMethodGao, T., Wang, Z., & Ji, Q. (2015) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.484 · Cited by 10
Huang, M., Mu, Z., & Zeng, H. (2015) · IET Image Processing · doi:10.1049/iet-ipr.2015.0329 · Cited by 10
Coupled-learning convolutional neural networks for object recognition
Xu, C., Yang, J. J., & Gao, J. (2017) · Multimedia Tools and Applications · doi:10.1007/s11042-017-5262-0 · Cited by 10
“Although the classification errors of Co-CNN models on ImageNet dataset are higher than the popular ResNet model, our proposed Co-CNN models can outperform its baselines: sparse coding [1], SIFT with fisher vector coding [28] and Alexnet [15] model.”
MethodSimilarity Embedding Networks for Robust Human Activity Recognition
Li, C., Tong, C. L., Niu, D., Jiang, B., Zuo, X. M., Cheng, L., Xiong, J., & Yang, J. (2021) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3448021 · Cited by 14
“Applications of learning to hash include large-scale object retrieval [13], image classification [32], and detection [35].”
BackgroundScalable Mobile Visual Classification by Kernel Preserving Projection Over High-Dimensional Features
Su, Y.-C., Chiu, T.-H., Kuo, Y.-H., Yeh, C.-Y., & Hsu, W. H. (2014) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2014.2322337 · Cited by 9
“Besides dimension reduction, a feature compression method–product quantization (PQ) [10]–is also adopted in large scale visual recognition.”
BackgroundMethodConvolutional Neural Network Applied to Tree Species Identification Based on Leaf Images
Minowa, Y., & Nagasaki, Y. (2020) · Journal of Forest Planning · doi:10.20659/jfp.2020.001 · Cited by 9
Utilizing geo-referenced imagery for systematic social observation of neighborhood disorder
Snaphaan, T., & Hardyns, W. (2021) · Computers Environment and Urban Systems · doi:10.1016/j.compenvurbsys.2021.101691 · Cited by 22
Online User Modeling for Interactive Streaming Image Classification
Hu, J., Sun, Z., Li, B., Yang, K., & Li, D. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-51814-5_25 · Cited by 9
Fried Binary Embedding: From High-Dimensional Visual Features to High-Dimensional Binary Codes
Hong, W., & Yuan, J. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2846670 · Cited by 9
“For the first challenge, it has been noticed that for input feature vector of dimensionality d , the length b of binary codes required to achieve reasonable accuracy is usually O(d) [10]–[15].”
BackgroundBanda, J. M., Schuh, M. A., Angryk, R. A., Pillai, K. G., & McInerney, P. (2013) · Advances in intelligent systems and computing · doi:10.1007/978-3-319-01863-8_17 · Cited by 9
An Efficient Image Categorization Method With Insufficient Training Samples
Lin, L., Liu, B., Zheng, X. W., & Xiao, Y. (2020) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2020.3011165 · Cited by 10
“For example, the shadow model [15] with the SIFT feature has an accuracy rate of about 16.7% on the Imagenet dataset.”
MethodInfluentialRepercussion of Image Compression on Satellite Image Classification using Deep Learning Models
Hossain, M. J., Barkatullah, M., Monir, M. F., & Ahmed, T. (2023) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2023-fall60731.2023.10333668 · Cited by 8
“(Jorge Sanchez and Florent Perronnin, 2011) [18] investigates image classification in large environments, investigating the effects of training set size and signature dimensionality, proposing lossy compression techniques for storage reduction without sacrificing accuracy, achieving significant…”
BackgroundBenchmarking large-scale Fine-Grained Categorization
Angelova, A., & Long, P. M. (2014) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2014.6836056 · Cited by 10
“[23].”
Negrel, R., Picard, D., & Gosselin, P.-H. (2014) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2014.7025444 · Cited by 10
“However signatures that perform well are very large [10], and the storage cost becomes pro-hibitive for web scale datasets.”
BackgroundLatent Task Adaptation with Large-Scale Hierarchies
Jia, Y., & Darrell, T. (2013) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2013.260 · Cited by 8
“Recent efforts on learning better classifiers, such as the ones presented in [21, 13] could be seamlessly incorporated into our learning framework for general performance increases.”
BackgroundHashing Graph Convolution for Node Classification
Zhao, W., Cui, Z., Xu, C., Li, C., Zhang, T., & Yang, J. J. (2019) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3357384.3357922 · Cited by 8
“Due to similarity preserving [33], learning to hash has been applied to a wide-range of applications such as large scale object retrieval [12], image classification [25] and detection [30], and so on.”
BackgroundIMAGE RECOGNITION USING MACHINE LEARNING
Patil, A. (2021) · International Journal of Engineering Applied Sciences and Technology · doi:10.33564/ijeast.2021.v06i01.027 · Cited by 8
Incremental Parallel Support Vector Machines for Classifying Large-Scale Multi-class Image Datasets
Do, T., & Tran-Nguyen, M.-T. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-48057-2_2 · Cited by 9
SRNN: Self-regularized neural network
Xu, C., Yang, J. J., Gao, J., Lai, H., & Yan, S. (2017) · Neurocomputing · doi:10.1016/j.neucom.2017.07.051 · Cited by 9
Li, D., & Chuah, M. C. (2015) · ACM Multimedia Systems Conference (MMSys) · doi:10.1145/2713168.2713172 · Cited by 7
“Various compression and encoding methods are proposed to the BOVW framework [13][17][37][21] to further reduce the memory impact.”
MethodAnalyzing sedentary behavior in life-logging images
Moghimi, M. K., Wu, W., Chen, J. M., Godbole, S. V., Marshall, S. J., Kerr, J., & Belongie, S. (2014) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2014.7025202 · Cited by 7
Liu, X., Nie, X., Xi, X., Zhu, L., & Yin, Y. (2019) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3357384.3357915 · Cited by 7
Deep Learning Based Image Classification Using Small VGG Net Architecture
Paulchamy, B., Maharajan, K., & Srikanteswara, R. (2022) · IEEE Mysore Sub Section International Conference (MysuruCon) · doi:10.1109/mysurucon55714.2022.9972441 · Cited by 7
“They're made up of neurons, which are the fundamental computational units in neural networks [10].”
BackgroundLearning Compact Visual Attributes for Large-Scale Image Classification
Su, Y., & Jurie, F. (2012) · Lecture notes in computer science · doi:10.1007/978-3-642-33885-4_6 · Cited by 7
“The quantization based techniques were also proposed to compress the high dimensional image signatures ( e.g. [ 11–13]).”
MethodInfluentialClassification of Occluded Objects Using Fast Recurrent Processing
Yılmaz, Ö. (2015) · International Conference on Machine Learning and Applications · doi:10.1109/icmla.2015.149 · Cited by 7
“Artificially expanding the data set using image transformations have been successfully used for reducing overfitting [72, 73, 74].”
MethodPandey, S., Khanna, P., & Yokota, H. (2016) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2016.04.013 · Cited by 8
Local Patch Vectors Encoded by Fisher Vectors for Image Classification
Chen, S., Liu, H., Zeng, X., Qian, S., Wei, W., Wu, G., & Duan, B. (2018) · Information · doi:10.3390/info9020038 · Cited by 7
“The second are hashing based [33], which transform a real-valued feature vector into a shorter binary string.”
BackgroundEnhancing Training Collections for Image Annotation: An Instance-Weighted Mixture Modeling Approach
Sawant, N., Wang, J. Z., & Li, J. (2013) · IEEE Transactions on Image Processing · doi:10.1109/tip.2013.2262289 · Cited by 6
Three Guidelines of Online Learning for Large-Scale Visual Recognition
Ushiku, Y., Hidaka, M., & Harada, T. (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.2014.457 · Cited by 7
“By virtue of recent advances in computer science and because of the culture of sharing of multimedia information such as photographs, vast quantities of labeled images have been used for visual recognition [14, 26, 34, 39].”
BackgroundMethodEfficient search over incomplete knowledge graphs in binarized embedding space
Wang, M., Chen, W., Wang, S., Jiang, Y., Yao, L., & Qi, G. (2021) · Future Generation Computer Systems · doi:10.1016/j.future.2021.04.006 · Cited by 6
Graphical Gaussian Vector for Image Categorization
Harada, T., & Kuniyoshi, Y. (2012) · Cited by 6
Ji, Z. (2013) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2013.74 · Cited by 6
“For an example of SVM, learning algorithms of Stochastic Gradient Descent (SGD) [2, 31] and Averaged SGD [22] are developed to train one-againstall SVMs incrementally and in parallel for a large amount of data.”
BackgroundMethodInfluentialSarowar, M. G., Razzak, M. A., & Fuad, M. A. A. (2019) · IEEE International Conference on Advanced Computing (IACC) · doi:10.1109/iacc48062.2019.8971585 · Cited by 5
“For the related problem of object recognition manually engineered representations in conjunction with shallow discriminatively trained models have been among the best performing paradigms [3].”
MethodA Modification-Free Steganography Algorithm Based on Image Classification and CNN
Wu, J., Zhang, Y., Luo, C. W., Yuan, L., & Shen, X. K. (2021) · International Journal of Digital Crime and Forensics · doi:10.4018/ijdcf.20210501.oa4 · Cited by 5
“We used Rectified Linear Units (ReLU) as neurons (Hinton et al., n.d.; Sánchez, 2011).”
MethodImage Classification Based on Convolutional Neural Network
Prassanna, P. L., Sandeep, S., Rao, K. V., Sasidhar, T. T., Lavanya, D. R., Deepthi, G., SriLakshmi, N. V., Mounika, P., & Govardhani, U. (2022) · Lecture notes on data engineering and communications technologies · doi:10.1007/978-981-16-6605-6_64 · Cited by 5
Kawano, Y., & Yanai, . (2014) · IPSJ Transactions on Computer Vision and Applications · doi:10.2197/ipsjtcva.6.83 · Cited by 6
“Note that we perform training of linear classifiers using PQ in the same as [10].”
BackgroundMethodInfluentialRecognition and Classification of Rotorcraft by Micro-Doppler Signatures Using Deep Learning
Liu, Y., & Liu, J. (2018) · Lecture notes in computer science · doi:10.1007/978-3-319-93698-7_11 · Cited by 5
Song, X., Wang, B., Chen, G., & Jiang, S. (2019) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3343031.3350913 · Cited by 5
“Classical approaches [16, 30, 34] exploit the handcrafted features Unpublished working draft. classroom use is granted without fee provided Not for distribution. that copies are not made or Permission to make digital or hard copies of all or part of this work for personal or distributed for profit…”
BackgroundLi, T., Alipour, M., Donaldson, B. M., & Harris, D. K. (2021) · Transportation Research Record Journal of the Transportation Research Board · doi:10.1177/03611981211031904 · Cited by 5
Hierarchical Decomposition of Large Deep Networks
Chennupati, S., Sah, S., Nooka, S. P., & Ptucha, R. (2016) · Electronic Imaging · doi:10.2352/issn.2470-1173.2016.19.coimg-152 · Cited by 6
“2010-[23] 2011-[24] 2012- [5] 2013-[2] 2014-[7] [3] 2015-[26] Top-5 error on ImageNet”
BackgroundOptimal feature-level fusion and layered k-support vector machine for spoofing face detection
Kavitha, P., & Vijaya, K. (2018) · Multimedia Tools and Applications · doi:10.1007/s11042-018-5877-9 · Cited by 5
“In many of the computer vision applications [42], extracting dense features ensures more effectiveness.”
Locally Enhanced Self-Attention: Combining Self-Attention and Convolution as Local and Context TermsPreprint
Yang, C., Qiao, S., Kortylewski, A., & Yuille, A. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.05637 · Cited by 7
Compact representation for large-scale unconstrained video analysis
Wang, S., Pan, P., Long, G., Chen, W., Li, X., & Sheng, Q. Z. (2015) · World Wide Web · doi:10.1007/s11280-015-0354-0 · Cited by 4
“In literature, feature compression is divided into three groups, including Product Quantization (PQ) [10, 29], hashing-based methods [24] and dimension reduction [4, 21, 26].”
On the use of convolutional neural networks for graphical model-based human pose estimation
Vu, H. T., Cheng, E., Wilkinson, R. H., & Lech, M. (2017) · International Conference on Recent Advances in Signal Processing, Telecommunications & Computing (SigTelCom) · doi:10.1109/sigtelcom.2017.7849801 · Cited by 4
“BACKGROUND CNN -based approaches for object classification and recognition [2] [23] [24] [25] obtained significantly higher classification accuracy compared to traditional methods [3] [26].”
ResultGeological mapping in the age of artificial intelligence
(2022) · Geoscientist · doi:10.1144/geosci2022-023 · Cited by 6
A Degradation Type Adaptive and Deep CNN-Based Image Classification Model for Degraded Images
Liu, H., Wang, W., Liu, H., Yi, S., Yonghao, Y., & Yao, X. (2023) · Computer Modeling in Engineering & Sciences · doi:10.32604/cmes.2023.029084 · Cited by 9
Optimization of Robust Loss Functions for Weakly-Labeled Image Taxonomies: An ImageNet Case Study
McAuley, J., Arnau Ramisa, A., & Caetano, T. S. (2011) · Lecture notes in computer science · doi:10.1007/978-3-642-23094-3_26 · Cited by 5
Image annotation with incomplete labelling by modelling image specific structured loss
Xu, X., Shimada, A., Nagahara, H., Taniguchi, R., & He, L. (2015) · IEEJ Transactions on Electrical and Electronic Engineering · doi:10.1002/tee.22190 · Cited by 5
RGB-D scene classification via heterogeneous model fusion
Liu, X., Wang, X., & Jiang, S. (2016) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2016.7532407 · Cited by 4
“Previous works[1, 2, 3, 4] mainly exploit the lowlevel visual descriptors.”
MethodMultiple feature fusion based image classification using a non-biased multi-scale kernel machine
Wang, H., Fu, G., Cai, Y., & Wang, S. (2015) · International Conference on Fuzzy Systems and Knowledge Discovery (FSKD) · doi:10.1109/fskd.2015.7382027 · Cited by 3
“INTRODUCTION As an important branch of pattern recognition, image classification([1]) is a research focus, which needs a comprehensive application of discipline knowledge in computer vision, pattern analysis and image understanding.”
BackgroundDoan, T.-N., Do, T., & Poulet, F. (2013) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2013.6707121 · Cited by 4
“The prominent works are proposed in [8], [9], [21], [22] where the data are first transformed by a nonlinear mapping induced by a particular kernel and then linear classifier is trained in the resulting space.”
Large Scale Metric Learning for Distance-Based Image Classification
Mensink, T., Verbeek, J. J., Perronnin, F., Csurka, G., Mensink, T., Verbeek, J., Perronnin, F., Csurka, G., Metric, L. S., Mensink, T., et al. (2012) · Cited by 3
Self-citationScalable image annotation using a product compressive sampling approach
Maronidis, A., Chatzilari, E., Nikolopoulos, S., & Kompatsiaris, I. Y. (2015) · IEEE International Conference on Data Science and Advanced Analytics (DSAA) · doi:10.1109/dsaa.2015.7344901 · Cited by 3
“decomposition scheme, in PCS the classification is performed in the reduced space, while on the contrary, in PQ the reduced data need to be firstly decompressed [28] adding an extra computational burden.”
MethodLearning to Recognize Hand-Held Objects from Scratch
Li, X., Jiang, S., Lv, X., & Chen, C. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-48896-7_52 · Cited by 3
“Depth information are available in HOD, so we cascade the two kinds of features and use RBF kernel which could perform non-linear transformations of the original input features in SVM.”
BackgroundInfluentialVideo Similarity Measurement and Search
Bekhet, S., Hassaballah, M., Ahmed, A. A., & Ahmed, A. H. (2018) · Studies in computational intelligence · doi:10.1007/978-3-030-03000-1_4 · Cited by 3
Guo, W., Guo, R., Du, X., Zhao, Y., Li, X., & Toe, T. T. (2022) · International Conference on Robotics and Computer Vision (ICRCV) · doi:10.1109/icrcv55858.2022.9953225 · Cited by 3
Binarized network embedding with community structural information
Liu, Y., Liu, J., Wang, Z., Wang, X., Zhang, F., & Xiao, Z. (2022) · Information Sciences · doi:10.1016/j.ins.2022.09.055 · Cited by 2
Metadata
- Authors
- Sanchez, J., & Perronnin, F.
- Affiliations
- Xerox (France) (FR)
- Venue
- CVPR 2011
- Publisher
- IEEE
- Published
- 2011-06
- Pages
- 1665-1672
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W2053229256
- Semantic Scholar
- eefcc7bcc05436dac9881acb4ff4e4a0b730e175
- Cited by
- 272
- References
- 34
- Citations by year
- 2026: 3 · 2025: 5 · 2024: 7 · 2023: 6 · 2022: 11 · 2021: 19 · 2020: 10 · 2019: 23 · 2018: 17 · 2017: 23 · 2016: 21 · 2015: 46 · 2014: 30 · 2013: 30 · 2012: 18
- FWCI
- 19.28
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Image and Video Retrieval Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Image Retrieval and Classification Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Keywords
- image classification, large-scale image classification, signature compression, lossy compression, dimensionality reduction, training data size, storage efficiency, scalable training, ImageNet, Flickr images
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.