DOC PREVIEW
UIUC CS 543 - Image Categorization

This preview shows page 1-2-3-24-25-26-27-49-50-51 out of 51 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 51 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

Image CategorizationLast classesToday’s class: categorizationSlide 4Slide 5Part 1: Image featuresGeneral Principles of RepresentationImage Representations: HistogramsSlide 12Slide 13Computing histogram distanceHistograms: Implementation issuesWhat kind of things do we compute histograms of?Slide 18Image Categorization: Bag of WordsBut what about layout?Spatial pyramidPart 2: ClassifiersLearning a classifierMany classifiers to choose fromNo Free Lunch TheoremBias-Variance Trade-offBias and VarianceChoosing the trade-offEffect of Training SizeHow to measure complexity?How to reduce variance?The perfect classification algorithmGenerative vs. Discriminative ClassifiersGenerative Classifier: Naïve BayesUsing Naïve BayesClassifiers: Logistic RegressionUsing Logistic RegressionClassifiers: Linear SVMSlide 48Slide 49Classifiers: Kernelized SVMUsing SVMsClassifiers: Decision TreesEnsemble Methods: BoostingBoosted Decision TreesUsing Boosted Decision TreesK-nearest neighbor1-nearest neighbor3-nearest neighbor5-nearest neighborUsing K-NNClustering (unsupervised)What to remember about classifiersNext classSome Machine Learning ReferencesImage CategorizationComputer VisionCS 543 / ECE 549 University of IllinoisDerek Hoiem03/11/10Last classes•Object recognition: localizing an object instance in an image•Face recognition: matching one face image to anotherToday’s class: categorization•Overview of image categorization•Representation–Image histograms•Classification–Important concepts in machine learning–What the classifiers are and when to use themImage CategorizationTraining LabelsTraining ImagesClassifier TrainingTrainingImage FeaturesTrained ClassifierImage CategorizationTraining LabelsTraining ImagesClassifier TrainingTrainingImage FeaturesImage FeaturesTestingTest ImageTrained ClassifierTrained ClassifierOutdoorPredictionPart 1: Image featuresTraining LabelsTraining ImagesClassifier TrainingTrainingImage FeaturesTrained ClassifierGeneral Principles of Representation•Coverage–Ensure that all relevant info is captured•Concision–Minimize number of features without sacrificing coverage•Directness–Ideal features are independently useful for predictionImage IntensitySpace Shuttle Cargo BayImage Representations: HistogramsGlobal histogram•Represent distribution of features–Color, texture, depth, …Images from Dave KauchakImage Representations: Histograms•Joint histogram–Requires lots of data–Loss of resolution to avoid empty binsImages from Dave KauchakMarginal histogram•Requires independent features•More data/bin than joint histogramHistogram: Probability or count of data in each binEASE Truss AssemblySpace Shuttle Cargo BayImage Representations: HistogramsImages from Dave KauchakClusteringUse the same cluster centers for all imagesComputing histogram distanceChi-squared Histogram matching distanceKmjijijimhmhmhmhhh122)()()]()([21),( Kmjijimhmhhh1)(),(min1),histint(Histogram intersection (assuming normalized histograms)Cars found by color histogram matching using chi-squaredHistograms: Implementation issuesFew BinsNeed less dataCoarser representationMany BinsNeed more dataFiner representation•Quantization–Grids: fast but only applicable with few dimensions–Clustering: slower but can quantize data in higher dimensions•Matching–Histogram intersection or Euclidean may be faster–Chi-squared often works better–Earth mover’s distance is good for when nearby bins represent similar valuesWhat kind of things do we compute histograms of?•Color•Texture (filter banks or HOG over regions)L*a*b* color space HSV color spaceWhat kind of things do we compute histograms of?•Histograms of gradient•Visual wordsSIFT – Lowe IJCV 2004Image Categorization: Bag of WordsTraining1. Extract keypoints and descriptors for all training images2. Cluster descriptors3. Quantize descriptors using cluster centers to get “visual words”4. Represent each image by normalized counts of “visual words” 5. Train classifier on labeled examples using histogram values as featuresTesting1. Extract keypoints/descriptors and quantize into visual words2. Compute visual word histogram3. Compute label or confidence using classifierBut what about layout?All of these images have the same color histogramSpatial pyramidCompute histogram in each spatial binPart 2: ClassifiersTraining LabelsTraining ImagesClassifier TrainingTrainingImage FeaturesTrained ClassifierLearning a classifier•Given some set features with corresponding labels, learn a function to predict the labels from the featuresx xxxxxxxooooox2x1Many classifiers to choose from•SVM•Neural networks•Naïve Bayes•Bayesian network•Logistic regression•Randomized Forests•Boosted Decision Trees•K-nearest neighbor•RBMs•Etc.Which is the best one?No Free Lunch TheoremBias-Variance Trade-ofMSE = bias2 + varianceBias and VarianceMany training examplesFew training examplesComplexityLow BiasHigh VarianceHigh BiasLow VarianceTest ErrorError = bias2 + varianceChoosing the trade-of•Need validation set•Validation set not same as test setTraining errorTest errorComplexityLow BiasHigh VarianceHigh BiasLow VarianceErrorEfect of Training SizeTestingTrainingNumber of Training ExamplesErrorGeneralization ErrorFixed classifierHow to measure complexity?•VC dimensionTraining error + Upper bound on generalization errorN: size of training seth: VC dimension: 1-probabilityHow to reduce variance?•Choose a simpler classifier•Regularize the parameters•Get more training dataThe perfect classification algorithm•Objective function: solves what you want to solve•Parameterization: makes assumptions that fit the problem•Regularization: right level of regularization for amount of training data•Training algorithm: can find parameters that maximize objective on training set•Inference algorithm: can solve for objective function in evaluationGenerative vs. Discriminative ClassifiersGenerative•Training–Maximize joint likelihood of data and labels–Assume (or learn) probability distribution and dependency structure–Can impose priors•Testing–P(y=1, x) / P(y=0, x) > t?•Examples–Foreground/background GMM –Naïve Bayes classifier–Bayesian networkDiscriminative•Training–Learn to directly predict the labels from the data–Assume form of boundary–Margin maximization or parameter regularization•Testing–f(x) > t ; e.g., wTx > t•Examples–Logistic


View Full Document
Download Image Categorization
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Image Categorization and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Image Categorization 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?