Data Driven Visual Recognition

This thesis is mostly about supervised visual recognition problems. Based on a general definition of categories, the contents are divided into two parts: one which models categories and one which is not category based. We are interested in data driven solutions for both kinds of problems. In the cat...

Full description

Bibliographic Details
Main Author: Aghazadeh, Omid
Format: Doctoral Thesis
Language:English
Published: KTH, Datorseende och robotik, CVAP 2014
Subjects:
Online Access:http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-145865
http://nbn-resolving.de/urn:isbn:978-91-7595-197-3
id ndltd-UPSALLA1-oai-DiVA.org-kth-145865
record_format oai_dc
spelling ndltd-UPSALLA1-oai-DiVA.org-kth-1458652017-02-23T05:26:58ZData Driven Visual RecognitionengAghazadeh, OmidKTH, Datorseende och robotik, CVAP2014Visual RecognitionData DrivenSupervised LearningMixture ModelsNon-Parametric ModelsCategory RecognitionNovelty DetectionThis thesis is mostly about supervised visual recognition problems. Based on a general definition of categories, the contents are divided into two parts: one which models categories and one which is not category based. We are interested in data driven solutions for both kinds of problems. In the category-free part, we study novelty detection in temporal and spatial domains as a category-free recognition problem. Using data driven models, we demonstrate that based on a few reference exemplars, our methods are able to detect novelties in ego-motions of people, and changes in the static environments surrounding them. In the category level part, we study object recognition. We consider both object category classification and localization, and propose scalable data driven approaches for both problems. A mixture of parametric classifiers, initialized with a sophisticated clustering of the training data, is demonstrated to adapt to the data better than various baselines such as the same model initialized with less subtly designed procedures. A nonparametric large margin classifier is introduced and demonstrated to have a multitude of advantages in comparison to its competitors: better training and testing time costs, the ability to make use of indefinite/invariant and deformable similarity measures, and adaptive complexity are the main features of the proposed model. We also propose a rather realistic model of recognition problems, which quantifies the interplay between representations, classifiers, and recognition performances. Based on data-describing measures which are aggregates of pairwise similarities of the training data, our model characterizes and describes the distributions of training exemplars. The measures are shown to capture many aspects of the difficulty of categorization problems and correlate significantly to the observed recognition performances. Utilizing these measures, the model predicts the performance of particular classifiers on distributions similar to the training data. These predictions, when compared to the test performance of the classifiers on the test sets, are reasonably accurate. We discuss various aspects of visual recognition problems: what is the interplay between representations and classification tasks, how can different models better adapt to the training data, etc. We describe and analyze the aforementioned methods that are designed to tackle different visual recognition problems, but share one common characteristic: being data driven. <p>QC 20140604</p>Doctoral thesis, comprehensive summaryinfo:eu-repo/semantics/doctoralThesistexthttp://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-145865urn:isbn:978-91-7595-197-3application/pdfinfo:eu-repo/semantics/openAccess
collection NDLTD
language English
format Doctoral Thesis
sources NDLTD
topic Visual Recognition
Data Driven
Supervised Learning
Mixture Models
Non-Parametric Models
Category Recognition
Novelty Detection
spellingShingle Visual Recognition
Data Driven
Supervised Learning
Mixture Models
Non-Parametric Models
Category Recognition
Novelty Detection
Aghazadeh, Omid
Data Driven Visual Recognition
description This thesis is mostly about supervised visual recognition problems. Based on a general definition of categories, the contents are divided into two parts: one which models categories and one which is not category based. We are interested in data driven solutions for both kinds of problems. In the category-free part, we study novelty detection in temporal and spatial domains as a category-free recognition problem. Using data driven models, we demonstrate that based on a few reference exemplars, our methods are able to detect novelties in ego-motions of people, and changes in the static environments surrounding them. In the category level part, we study object recognition. We consider both object category classification and localization, and propose scalable data driven approaches for both problems. A mixture of parametric classifiers, initialized with a sophisticated clustering of the training data, is demonstrated to adapt to the data better than various baselines such as the same model initialized with less subtly designed procedures. A nonparametric large margin classifier is introduced and demonstrated to have a multitude of advantages in comparison to its competitors: better training and testing time costs, the ability to make use of indefinite/invariant and deformable similarity measures, and adaptive complexity are the main features of the proposed model. We also propose a rather realistic model of recognition problems, which quantifies the interplay between representations, classifiers, and recognition performances. Based on data-describing measures which are aggregates of pairwise similarities of the training data, our model characterizes and describes the distributions of training exemplars. The measures are shown to capture many aspects of the difficulty of categorization problems and correlate significantly to the observed recognition performances. Utilizing these measures, the model predicts the performance of particular classifiers on distributions similar to the training data. These predictions, when compared to the test performance of the classifiers on the test sets, are reasonably accurate. We discuss various aspects of visual recognition problems: what is the interplay between representations and classification tasks, how can different models better adapt to the training data, etc. We describe and analyze the aforementioned methods that are designed to tackle different visual recognition problems, but share one common characteristic: being data driven. === <p>QC 20140604</p>
author Aghazadeh, Omid
author_facet Aghazadeh, Omid
author_sort Aghazadeh, Omid
title Data Driven Visual Recognition
title_short Data Driven Visual Recognition
title_full Data Driven Visual Recognition
title_fullStr Data Driven Visual Recognition
title_full_unstemmed Data Driven Visual Recognition
title_sort data driven visual recognition
publisher KTH, Datorseende och robotik, CVAP
publishDate 2014
url http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-145865
http://nbn-resolving.de/urn:isbn:978-91-7595-197-3
work_keys_str_mv AT aghazadehomid datadrivenvisualrecognition
_version_ 1718416395061952512