be Gentimentalmente 12 43 133 23 CD1 1 De Wurms made their 3.00 3.3 2.8 1.ET DAT 10000 100 25. NOT own. WWW Tom end Given three images in the database. we extract local features from each of them. Each feature is a t-by-3 vector, we can therefore denote each image by a N.by-3 matrix 143 4 3 3 41 42 Ti 12 13 3 2 3 3 2 2 2 2 2 2 42 T2 23 Glven a query image denoted the same way as above: 2 1 1 2 2 1 Use the Bag of Words (Bow) model to encode these images and compute the distances from the query image to the database images. Assume that the words are 3.00 2.25 2.50 V 1.67 4.00 1.67 1.00 4.00 3.00 2 we do NOT remove/dowweight feature vectors visual words that commonly occur in many images: 3) we take the Euclidean distance as the similarity measure ANSWER The Eldean distance from Qto it: (to 2 decimal places The Eldean distance from 0 to 12 to 2 decimal places The Bulldean distance from 0 to 13: (to 2 decimal places ature is al by vector we can therefore denote each image by a Nby amat the number of features that ingen ces from the query Amage to the database images. Assume that three visual words are pre dustered and offered in a sya matrix were each tow is the vector for one word cour in many images we take the Euclidean distante as the similarity measure. be Gentimentalmente 12 43 133 23 CD1 1 De Wurms made their 3.00 3.3 2.8 1.ET DAT 10000 100 25. NOT own. WWW Tom end Given three images in the database. we extract local features from each of them. Each feature is a t-by-3 vector, we can therefore denote each image by a N.by-3 matrix 143 4 3 3 41 42 Ti 12 13 3 2 3 3 2 2 2 2 2 2 42 T2 23 Glven a query image denoted the same way as above: 2 1 1 2 2 1 Use the Bag of Words (Bow) model to encode these images and compute the distances from the query image to the database images. Assume that the words are 3.00 2.25 2.50 V 1.67 4.00 1.67 1.00 4.00 3.00 2 we do NOT remove/dowweight feature vectors visual words that commonly occur in many images: 3) we take the Euclidean distance as the similarity measure ANSWER The Eldean distance from Qto it: (to 2 decimal places The Eldean distance from 0 to 12 to 2 decimal places The Bulldean distance from 0 to 13: (to 2 decimal places ature is al by vector we can therefore denote each image by a Nby amat the number of features that ingen ces from the query Amage to the database images. Assume that three visual words are pre dustered and offered in a sya matrix were each tow is the vector for one word cour in many images we take the Euclidean distante as the similarity measure
