We have developed a two-phase generative / discriminative learning procedure for the recognition of classes of objects and concepts in outdoor scenes. Our method uses both multiple types of object features and context within the image. The generative phase normalizes the description length of images, which can have an arbitrary number of extracted features of each type. In the discriminative phase, a classifier learns which images, as represented by this fixed-length description, contain the target object. We have tested the approach by comparing it to several other approaches in the literature and by experimenting with several different data sets and combinations of features. Our results, using color, texture, and structure features, show a significant improvement over previously published results in image retrieval. Using salient region features, we are competitive with recent results in object recognition. 1.

Classifying pictures into one of several semantic categories is a classical image understanding problem. In this paper, we present a stratified approach to both binary (outdoor-indoor) and multiple category of scene classification. We first learn mixture models for 20 basic classes of local image content based on color and texture information. Once trained, these models are applied to a test image, and produce 20 probability density response maps (PDRM) indicating the likelihood that each image region was produced by each class. We then extract some very simple features from those PDRMs, and use them to train a bagged LDA classifier for 10 scene categories. For this process, no explicit region segmentation or spatial context model are computed. To test this classification system, we created a labeled database of 1500 photos taken under very different environment and lighting conditions, using different cameras, and from 43 persons over 5 years. The classification rate of outdoor-indoor classification is 93.8%, and the classification rate for 10 scene categories is 90.1%. As a byproduct, local image patches can be contextually labeled into the 20 basic material classes by using Loopy Belief Propagation [33] as an anisotropic filter on PDRMs, producing an image-level segmentation if desired. 1

This dissertation addresses the problems of recognizing textures, objects, and scenes in photographs. We present approaches to these recognition tasks that combine salient local image features with spatial relations and effective discriminative learning techniques. First, we introduce a bag of features image model for recognizing textured surfaces under a wide range of transformations, including viewpoint changes and non-rigid deformations. We present results of a large-scale comparative evaluation indicating that bags of features can be effective not only for texture, but also for object categization, even in the presence of substantial clutter and intra-class variation. We also show how to augment the purely local image representation with statistical co-occurrence relations between pairs of nearby features, and develop a learning and classification framework for the task of classifying individual features in a multi-texture image. Next, we present a more structured alternative to bags of features for object recognition, namely, an image representation based on semi-local parts, or groups of features characterized by stable appearance and geometric layout. Semi-local parts are automatically learned from small sets of unsegmented, cluttered images. Finally, we present a global method for recognizing scene categories that works by partitioning the image into increasingly fine sub-regions and computing histograms of local features found inside each sub-region. The resulting spatial pyramid representation demonstrates significantly improved performance on challenging scene categorization tasks.

We have developed a two-phase generative / discriminative learning procedure for the recognition of classes of objects and concepts in outdoor scenes. Our method uses both multiple types of object features and context within the image. The generative phase normalizes the description length of images, which can have an arbitrary number of extracted features of each type. In the discriminative phase, a classifier learns which images, as represented by this fixed-length description, contain the target object. We have tested the approach by comparing it to several other approaches in the literature and by experimenting with several different data sets and combinations of features. Our results, using color, texture, and structure features, show a significant improvement over previously published results in image retrieval. Using salient region features, we are competitive with recent results in object recognition.

Abstract. The process of mixing labelled and unlabelled data is being recently studied in semi-supervision techniques. However, this is not the only scenario in which mixture of labelled and unlabelled data can be done. In this paper we propose a new problem we have called particularization and a way to solve it. We also propose a new technique for mixing labelled and unlabelled data. This technique relies in the combination of supervised and unsupervised processes competing for the classification of each data point. Encouraging results on improving the classification outcome are obtained on MNIST database. 1