Note 18 for ECE 582 with Professor Hao at UA
Note 18 for ECE 582 with Professor Hao at UA
Popular in Course
Popular in Department
This 9 page Class Notes was uploaded by an elite notetaker on Friday February 6, 2015. The Class Notes belongs to a course at University of Alabama - Tuscaloosa taught by a professor in Fall. Since its upload, it has received 18 views.
Reviews for Note 18 for ECE 582 with Professor Hao at UA
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 02/06/15
g Object Recognition Lecture 18 EE 482582 Computer Vision amp Digital Image Processing g Overview Classification Data Learning Supervised labeled training data Unsupervised unlabeled training data Decision Hypothesis Testing Distance based distance definition Euclidean Mahalanobis Correlation based Graph based hierarchical clustering Kernel based support vector machine A General Diagram for Pattern g Identification Systems F eature Model 1 39 39 Feature Feature raining lgt Igt Igt Extraction Clustering Feature Model K 1 Testing quotgt FeatUl39e lgt Feature lgt Hypothesis Emmetion Validation Tes ng lt2 The most important aspect of a pattern identi cation Unregistered Registered s stem is the selection of features Objects Objects Representation and Description Representation used to make the date useful to a computer further process description Description describes the region based on the chosen representation Example representation cgt boundary description cgt length of the boundary g Feature and Pattern Feature quantity measures of descriptions Pattern an arrangement of features Pattern class a set of patterns sharing some common properties Pattern recognition assign patterns to their respective classes A typical pattern is either A vector quantitative description A stringtree structural description g Classifiers Linear Classifiers Bayes classifier 20 s Fisher s linear discriminant 30 s Percepton 30 s Nonlinear Classifiers K mean Matlab statistics toolbox Decision tree Matlab statistics toolbox Neural networks 80 s Matlab neural network toolbox Bayesian networks 90 s Hidden Markov model 80 s Matlab HMM toolbox Support vector machine iArtificial Neural Networks Input Hidden Output nform anon ow 5 umd rechona Data 5 presented m mp0 layer Passed an m Hidden Layer wthls Passed an m Output layer nform anon 5 dwstrbuted Womanon procesSmg 5 paraHe Inlormahon gt An Example mm mmquot 0mmquot Q 1 x 025 05 x15 025 075 05 03775 Squashing 1 05 e Training the Network Learning Backpropagation Requires training set input output pairs Starts with small random weights Error is used to adjust weights supervised learning 9 Gradient descent on error landscape Supervised Learning Wallace lt 739 WINc Darwin miniquot mi Speaker Recognition Steve 043 026 David J 073 7 055 E 395 Face Recognition i w Targetl C1utter0 I Targeto C1utter1 g Eigen Face 1 Prepare a training set The faces constituting the training set T should be already prepared for processing Subtract the mean The average matrix A has to be calculated and subtracted from the original in T The results are stored in variable S Calculate the covariance matrix Calculate the eigenvectors and eigenvalues of this covariance matrix Choose the principal components g Hidden Markov Model Hidden Markov Model Training 1 Expectation amp 2 Maximization BaumWelch algorithm Evaluation forwardbackward algorithm Decoding viterbi algorithm Function Dynamic feature clusteringtesting Finite State Sequence Estimation Index ofevent Indexofstate o N A 0 on 5Y 1 0 1 5 2 0 2395 3390 3395 40 A four state HMM each with Number ofsamples two possible observations Gait Recognition g Matlab Statistics Toolbox Estimate a HMM model from observations and states trans emis hmmestimateseq states Estimate a HMM model from observations trans emis hmmtrainseq transguess emisguess Decode a state sequence out of the observation sequence and calculate the likelihood of the observation sequence generated by this model states likelihood hmmdecodeseq trans emis Unsupervised Learning 0 0s 04 03 02 01 0 0102 u 04 us Respmsesigiais C ounts El Probability Density 0s 04 03 02 01 0 0102 u 04 us Respmsesigiais
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'