IEEE Trans Pattern Anal Mach Intell - Distance-Based Image Classification: Generalizing to New Classes at Near Zero Cost.

Tópicos

{ learn(2355) train(1041) set(1003) }
{ imag(2830) propos(1344) filter(1198) }
{ can(981) present(881) function(850) }
{ perform(999) metric(946) measur(919) }
{ result(1111) use(1088) new(759) }
{ perform(1367) use(1326) method(1137) }
{ imag(1947) propos(1133) code(1026) }
{ control(1307) perform(991) simul(935) }
{ method(1219) similar(1157) match(930) }
{ concept(1167) ontolog(924) domain(897) }
{ algorithm(1844) comput(1787) effici(935) }
{ method(1557) propos(1049) approach(1037) }
{ data(2317) use(1299) case(1017) }
{ activ(1138) subject(705) human(624) }
{ data(1737) use(1416) pattern(1282) }
{ sequenc(1873) structur(1644) protein(1328) }
{ treatment(1704) effect(941) patient(846) }
{ problem(2511) optim(1539) algorithm(950) }
{ error(1145) method(1030) estim(1020) }
{ research(1085) discuss(1038) issu(1018) }
{ visual(1396) interact(850) tool(830) }
{ compound(1573) activ(1297) structur(1058) }
{ medic(1828) order(1363) alert(1069) }
{ cost(1906) reduc(1198) effect(832) }
{ group(2977) signific(1463) compar(1072) }
{ gene(2352) biolog(1181) express(1162) }
{ first(2504) two(1366) second(1323) }
{ analysi(2126) use(1163) compon(1037) }
{ implement(1333) system(1263) develop(1122) }
{ method(2212) result(1239) propos(1039) }
{ model(3404) distribut(989) bayesian(671) }
{ can(774) often(719) complex(702) }
{ inform(2794) health(2639) internet(1427) }
{ system(1976) rule(880) can(841) }
{ measur(2081) correl(1212) valu(896) }
{ imag(1057) registr(996) error(939) }
{ bind(1733) structur(1185) ligand(1036) }
{ featur(3375) classif(2383) classifi(1994) }
{ network(2748) neural(1063) input(814) }
{ imag(2675) segment(2577) method(1081) }
{ patient(2315) diseas(1263) diabet(1191) }
{ take(945) account(800) differ(722) }
{ studi(2440) review(1878) systemat(933) }
{ motion(1329) object(1292) video(1091) }
{ assess(1506) score(1403) qualiti(1306) }
{ surgeri(1148) surgic(1085) robot(1054) }
{ framework(1458) process(801) describ(734) }
{ chang(1828) time(1643) increas(1301) }
{ clinic(1479) use(1117) guidelin(835) }
{ extract(1171) text(1153) clinic(932) }
{ data(1714) softwar(1251) tool(1186) }
{ design(1359) user(1324) use(1319) }
{ model(2220) cell(1177) simul(1124) }
{ care(1570) inform(1187) nurs(1089) }
{ general(901) number(790) one(736) }
{ method(984) reconstruct(947) comput(926) }
{ search(2224) databas(1162) retriev(909) }
{ featur(1941) imag(1645) propos(1176) }
{ case(1353) use(1143) diagnosi(1136) }
{ howev(809) still(633) remain(590) }
{ data(3963) clinic(1234) research(1004) }
{ studi(1410) differ(1259) use(1210) }
{ risk(3053) factor(974) diseas(938) }
{ system(1050) medic(1026) inform(1018) }
{ import(1318) role(1303) understand(862) }
{ model(2341) predict(2261) use(1141) }
{ studi(1119) effect(1106) posit(819) }
{ blood(1257) pressur(1144) flow(957) }
{ spatial(1525) area(1432) region(1030) }
{ record(1888) medic(1808) patient(1693) }
{ health(3367) inform(1360) care(1135) }
{ model(3480) simul(1196) paramet(876) }
{ monitor(1329) mobil(1314) devic(1160) }
{ ehr(2073) health(1662) electron(1139) }
{ state(1844) use(1261) util(961) }
{ research(1218) medic(880) student(794) }
{ patient(2837) hospit(1953) medic(668) }
{ model(2656) set(1616) predict(1553) }
{ age(1611) year(1155) adult(843) }
{ signal(2180) analysi(812) frequenc(800) }
{ sampl(1606) size(1419) use(1276) }
{ data(3008) multipl(1320) sourc(1022) }
{ intervent(3218) particip(2042) group(1664) }
{ time(1939) patient(1703) rate(768) }
{ patient(1821) servic(1111) care(1106) }
{ use(2086) technolog(871) perceiv(783) }
{ health(1844) social(1437) communiti(874) }
{ structur(1116) can(940) graph(676) }
{ high(1669) rate(1365) level(1280) }
{ cancer(2502) breast(956) screen(824) }
{ use(976) code(926) identifi(902) }
{ use(1733) differ(960) four(931) }
{ drug(1928) target(777) effect(648) }
{ survey(1388) particip(1329) question(1065) }
{ estim(2440) model(1874) function(577) }
{ decis(3086) make(1611) patient(1517) }
{ process(1125) use(805) approach(778) }
{ activ(1452) weight(1219) physic(1104) }
{ method(1969) cluster(1462) data(1082) }
{ detect(2391) sensit(1101) algorithm(908) }

Resumo

We study large-scale image classification methods that can incorporate new classes and training images continuously over time at negligible cost. To this end we consider two distance-based classifiers, the k-nearest neighbor (k-NN) and nearest class mean (NCM) classifiers, and introduce a new metric learning approach for the latter. We also introduce an extension of the NCM classifier to allow for richer class representations. Experiments on the ImageNet 2010 challenge dataset, which contains over a million training images of thousand classes, show that, surprisingly, the NCM classifier compares favorably to the more flexible k-NN classifier. Moreover, the NCM performance is comparable to that of linear SVMs which obtain current state-of-the-art performance. Experimentally we study the generalization performance to classes that were not used to learn the metrics. Using a metric learned on 1,000 classes, we show results for the ImageNet-10K dataset which contains 10,000 classes, and obtain performance that is competitive with the current state-of-the-art, while being orders of magnitude faster. Furthermore, we show how a zero-shot class prior based on the ImageNet hierarchy can improve performance when few training images are available.

Resumo Limpo

studi largescal imag classif method can incorpor new class train imag continu time neglig cost end consid two distancebas classifi knearest neighbor knn nearest class mean ncm classifi introduc new metric learn approach latter also introduc extens ncm classifi allow richer class represent experi imagenet challeng dataset contain million train imag thousand class show surpris ncm classifi compar favor flexibl knn classifi moreov ncm perform compar linear svms obtain current stateoftheart perform experiment studi general perform class use learn metric use metric learn class show result imagenetk dataset contain class obtain perform competit current stateoftheart order magnitud faster furthermor show zeroshot class prior base imagenet hierarchi can improv perform train imag avail

Resumos Similares

Comput Math Methods Med - On multilabel classification methods of incompletely labeled biomedical text data. ( 0,886207065839178 )
IEEE Trans Neural Netw Learn Syst - A Kernel Classification Framework for Metric Learning. ( 0,885139385838796 )
IEEE Trans Image Process - Manifold regularized multitask learning for semi-supervised multilabel image classification. ( 0,847140017435657 )
Int J Neural Syst - Structurally enhanced incremental neural learning for image classification with subgraph extraction. ( 0,845680523759054 )
IEEE Trans Image Process - Hyperspectral image classification through bilayer graph-based learning. ( 0,827387798618412 )
IEEE Trans Image Process - Geodesic propagation for semantic labeling. ( 0,822999313580014 )
J Am Med Inform Assoc - Learning classification models with soft-label information. ( 0,820216569402708 )
Neural Comput - Reduction from cost-sensitive ordinal ranking to weighted binary classification. ( 0,817247723028802 )
Neural Comput - Computing sparse representations of multidimensional signals using Kronecker bases. ( 0,814989567064081 )
IEEE Trans Image Process - Joint segmentation of images and scanned point cloud in large-scale street scenes with low-annotation cost. ( 0,813188511663041 )
Neural Comput - Adaptive metric learning vector quantization for ordinal classification. ( 0,810897772991805 )
J. Comput. Biol. - Imbalanced class learning in epigenetics. ( 0,807093256712571 )
IEEE Trans Neural Netw Learn Syst - Adaptive Batch Mode Active Learning. ( 0,806100942450662 )
J Biomed Inform - Incremental Gaussian Discriminant Analysis based on Graybill and Deal weighted combination of estimators for brain tumour diagnosis. ( 0,803868725068155 )
IEEE Trans Image Process - Task-specific image partitioning. ( 0,803309717107224 )
IEEE Trans Image Process - Active learning for solving the incomplete data problem in facial age classification by the furthest nearest-neighbor criterion. ( 0,802011917046527 )
Int J Neural Syst - Online semi-supervised growing neural gas. ( 0,801654262287192 )
Int J Neural Syst - Aggregation of sparse linear discriminant analyses for event-related potential classification in brain-computer interface. ( 0,801220112039857 )
J Biomed Inform - Semi-supervised clinical text classification with Laplacian SVMs: an application to cancer case management. ( 0,799713272909187 )
Int J Neural Syst - Linear time relational prototype based learning. ( 0,798231066557816 )
IEEE Trans Image Process - A linear support higher-order tensor machine for classification. ( 0,78386116662823 )
J Med Syst - 3D similarity-dissimilarity plot for high dimensional data visualization in the context of biomedical pattern classification. ( 0,783412063529344 )
J Biomed Inform - Class proximity measures--dissimilarity-based classification and display of high-dimensional data. ( 0,782145915897397 )
Comput. Biol. Med. - Sparse Manifold Clustering and Embedding to discriminate gene expression profiles of glioblastoma and meningioma tumors. ( 0,780490635003893 )
IEEE Trans Pattern Anal Mach Intell - Good Practice in Large-Scale Learning for Image Classification. ( 0,780451105902677 )
J Am Med Inform Assoc - Active learning for clinical text classification: is it better than random sampling? ( 0,77717609463932 )
Neural Comput - Metacognitive learning in a fully complex-valued radial basis function neural network. ( 0,77431835727389 )
Int J Neural Syst - Span: spike pattern association neuron for learning spatio-temporal spike patterns. ( 0,77282629619303 )
Comput. Biol. Med. - Robust prediction of protein subcellular localization combining PCA and WSVMs. ( 0,772060602566344 )
IEEE Trans Image Process - Artistic image analysis using graph-based learning approaches. ( 0,76601225426749 )
IEEE Trans Pattern Anal Mach Intell - Label Consistent K-SVD: Learning A Discriminative Dictionary for Recognition. ( 0,76014755781451 )
J Biomed Inform - Multi-label classification of chronically ill patients with bag of words and supervised dimensionality reduction algorithms. ( 0,760091225240483 )
IEEE Trans Image Process - Improving Web image search by bag-based reranking. ( 0,759997390951369 )
J Biomed Inform - Portable automatic text classification for adverse drug reaction detection via multi-corpus training. ( 0,759047116757743 )
IEEE Trans Image Process - Self-supervised online metric learning with low rank constraint for scene categorization. ( 0,758600496958351 )
Neural Comput - Divergence-based vector quantization. ( 0,757420911604316 )
J. Comput. Biol. - Locally learning biomedical data using diffusion frames. ( 0,756030101453909 )
IEEE Trans Pattern Anal Mach Intell - Weakly Supervised Recognition of Daily Life Activities with Wearable Sensors. ( 0,755898545654074 )
Neural Comput - Online learning with (multiple) kernels: a review. ( 0,755208977009812 )
Neural Comput - Multiple spectral kernel learning and a gaussian complexity computation. ( 0,755142412255446 )
IEEE Trans Image Process - Unsupervised amplitude and texture classification of SAR images with multinomial latent model. ( 0,753986521526279 )
IEEE J Biomed Health Inform - Systematic Poisoning Attacks on and Defenses for Machine Learning in Healthcare. ( 0,752709278843609 )
IEEE Trans Pattern Anal Mach Intell - Feature Selection with Conjunctions of Decision Stumps and Learning from Microarray Data. ( 0,751934441161567 )
IEEE Trans Image Process - Design of non-linear kernel dictionaries for object recognition. ( 0,750156112938249 )
Comput Math Methods Med - Correlation kernels for support vector machines classification with applications in cancer data. ( 0,747984845774507 )
Neural Comput - Unsupervised learning of generative and discriminative weights encoding elementary image components in a predictive coding model of cortical function. ( 0,742252526434746 )
IEEE Trans Image Process - Structured max-margin learning for inter-related classifier training and multilabel image annotation. ( 0,741341925183145 )
J Biomed Inform - Applying active learning to assertion classification of concepts in clinical text. ( 0,737793490334534 )
IEEE Trans Pattern Anal Mach Intell - Learning Categories from Few Examples with Multi Model Knowledge Transfer. ( 0,732590706471208 )
IEEE Trans Image Process - Supervised ordering in IRp: application to morphological processing of hyperspectral images. ( 0,732502328707971 )
IEEE Trans Image Process - Multiview Hessian regularization for image annotation. ( 0,729114348914347 )
IEEE Trans Image Process - Multiple-kernel, multiple-instance similarity features for efficient visual object detection. ( 0,72442655194799 )
IEEE Trans Pattern Anal Mach Intell - The Effect of Model Misspecification on Semi-Supervised Classification. ( 0,723399277668089 )
Comput Methods Programs Biomed - Auto-adaptive robot-aided therapy using machine learning techniques. ( 0,721648222316187 )
Neural Comput - Representing objects, relations, and sequences. ( 0,718339873965629 )
J Biomed Inform - Learning classification models from multiple experts. ( 0,714558393341253 )
Comput Methods Programs Biomed - Multistage approach for clustering and classification of ECG data. ( 0,713594841819797 )
Med Decis Making - The Impact of Oversampling with SMOTE on the Performance of 3 Classifiers in Prediction of Type 2 Diabetes. ( 0,711927717261711 )
BMC Med Inform Decis Mak - Learning to improve medical decision making from imbalanced data without a priori cost. ( 0,711660342416532 )
IEEE Trans Image Process - Learning discriminative dictionary for group sparse representation. ( 0,709133827772346 )
IEEE Trans Neural Netw Learn Syst - An efficient topological distance-based tree kernel. ( 0,708620032446598 )
IEEE Trans Neural Netw Learn Syst - ML-Tree: a tree-structure-based approach to multilabel learning. ( 0,707206212756038 )
J Biomed Inform - Active learning strategies for the deduplication of electronic patient data using classification trees. ( 0,706629285224584 )
AMIA Annu Symp Proc - Comparison and combination of several MeSH indexing approaches. ( 0,704737153728179 )
Artif Intell Med - A classifier ensemble approach for the missing feature problem. ( 0,70456806439309 )
IEEE Trans Neural Netw Learn Syst - Partially shared latent factor learning with multiview data. ( 0,701034987816609 )
Comput Methods Programs Biomed - Modified CC-LR algorithm with three diverse feature sets for motor imagery tasks classification in EEG based brain-computer interface. ( 0,698111630368018 )
Comput Methods Programs Biomed - Comparison of machine learning methods for classifying aphasic and non-aphasic speakers. ( 0,697688821409815 )
Artif Intell Med - A fuzzy-based data transformation for feature extraction to increase classification performance with small medical data sets. ( 0,69458393119742 )
IEEE Trans Image Process - Incremental training of a detector using online sparse eigendecomposition. ( 0,694515344152968 )
J Chem Inf Model - Training based on ligand efficiency improves prediction of bioactivities of ligands and drug target proteins in a machine learning approach. ( 0,694388496401754 )
Neural Comput - Large margin low rank tensor analysis. ( 0,693963572414446 )
IEEE Trans Pattern Anal Mach Intell - Representation Learning: A Review and New Perspectives. ( 0,693259818907501 )
Neural Comput - Incremental learning by message passing in hierarchical temporal memory. ( 0,691229218298307 )
IEEE Trans Pattern Anal Mach Intell - Facial Age Estimation by Learning from Label Distributions. ( 0,687930361912777 )
J Chem Inf Model - Modeling and benchmark data set for the inhibition of c-Jun N-terminal kinase-3. ( 0,687604470695713 )
IEEE Trans Image Process - Saliency and gist features for target detection in satellite images. ( 0,687326866974025 )
IEEE Trans Image Process - Data-dependent hashing based on p-stable distribution. ( 0,684832621854396 )
IEEE Trans Image Process - Fast semantic diffusion for large-scale context-based image and video annotation. ( 0,684802516034892 )
IEEE Trans Image Process - Dimensionality reduction for registration of high-dimensional data sets. ( 0,683776883257353 )
IEEE Trans Pattern Anal Mach Intell - Feature Selection and Kernel Learning for Local Learning-Based Clustering. ( 0,682842187054283 )
Int J Comput Assist Radiol Surg - Statistical shape model of a liver for autopsy imaging. ( 0,675032495645201 )
IEEE Trans Image Process - Decomposition-based transfer distance metric learning for image classification. ( 0,675002616237151 )
IEEE Trans Pattern Anal Mach Intell - Latent Dirichlet Allocation Models for Image Classification. ( 0,674749297365855 )
J Biomed Inform - Supervised methods for symptom name recognition in free-text clinical records of traditional Chinese medicine: an empirical study. ( 0,673409325951497 )
IEEE Trans Image Process - Local learned dictionaries optimized to edge orientation for inverse halftoning. ( 0,670254427345593 )
IEEE Trans Neural Netw Learn Syst - Ordinal Distance Metric Learning for Image Ranking. ( 0,66995615811164 )
J Chem Inf Model - Classifying large chemical data sets: using a regularized potential function method. ( 0,669793018326729 )
Int J Neural Syst - Epileptic EEG classification based on kernel sparse representation. ( 0,667781487685237 )
IEEE Trans Image Process - Contextual kernel and spectral methods for learning the semantics of images. ( 0,66759889561427 )
AMIA Annu Symp Proc - Sample-efficient learning with auxiliary class-label information. ( 0,666840549146264 )
IEEE Trans Image Process - Enhancing training collections for image annotation: an instance-weighted mixture modeling approach. ( 0,665938970981471 )
IEEE Trans Image Process - Cross-Device Automated Prostate Cancer Localization With Multiparametric MRI. ( 0,663859511738919 )
Neural Comput - Mismatched training and test distributions can outperform matched ones. ( 0,663660224687418 )
Neural Comput - Adaptive multiclass classification for brain computer interfaces. ( 0,660482882755325 )
IEEE Trans Pattern Anal Mach Intell - A Bag-of-Features Framework to Classify Time Series. ( 0,657974583790084 )
J Chem Inf Model - Note on naive Bayes based on binary descriptors in cheminformatics. ( 0,656199829009147 )
Methods Inf Med - Probability machines: consistent probability estimation using nonparametric learning machines. ( 0,654117616879504 )
IEEE Trans Pattern Anal Mach Intell - Learning Hierarchical Features for Scene Labeling. ( 0,654055373317667 )
IEEE Trans Image Process - Fast bilateral filter with arbitrary range and domain kernels. ( 0,652496021636094 )