Chad Carson

Region-based image querying

Retrieving images from large and varied collections using image content as a key is a challenging and important problem. In this paper, we present a new image representation which provides a transformation from the raw pixel data to a small set of localized coherent regions in color and texture space. This so-called “blobworld” representation is based on segmentation using the expectation-maximization algorithm on combined color and texture features. The texture features we use for the segmentation arise from a new approach to texture description and scale selection. We describe a system that uses the blobworld representation to retrieve images. An important and unique aspect of the system is that, in the context of similarity-based querying, the user is allowed to view the internal representation of the submitted image and the query results. Similar systems do not offer the user this view into the workings of the system; consequently, the outcome of many queries on these systems can be quite inexplicable, despite the availability of knobs for adjusting the similarity metric

Finding Pictures of Objects in Large Collections of Images

Retrieving images from very large collections, using image content as a key, is becoming an important problem. Users prefer to ask for pictures using notions of content that are strongly oriented to the presence of abstractly defined objects. Computer programs that implement these queries automatically are desirable, but are hard to build because conventional object recognition techniques from computer vision cannot recognize very general objects in very general contexts. This paper describes our approach to object recognition, which is structured around a sequence of increasingly specialized grouping activities that assemble coherent regions of image that can be shown to satisfy increasingly stringent constraints. The constraints that are satisfied provide a form of object classification in quite general contexts. This view of recognition is distinguished by: far richer involvement of early visual primitives, including color and texture; hierarchical grouping and learning strategies in the classification process; the ability to deal with rather general objects in uncontrolled configurations and contexts. We illustrate these properties with four case-studies: one demonstrating the use of color and texture descriptors; one showing how trees can be described by fusing texture and geometric properties; one learning scenery concepts using grouped features; and one showing how this view of recognition yields a program that can tell, quite accurately, whether a picture contains naked people or not.

Creating a Customized Access Method for Blobworld

We present the design and analysis of a customized access method for the content-based image retrieval system, Blobworld. Using the amdb access method analysis tool, we analyzed three existing multidimensional access methods to support nearest neighbor search in the context of the Blobworld application. Based on this analysis, we propose several variants of the R-tree, tailored to address the problems the analysis revealed. We implemented the access methods we propose in the Generalized Search Trees (GiST) framework and analyzed them. We found that two of our access methods have better performance characteristics for the Blobworld application than any of the traditional multi-dimensional access methods we examined. Based on this experience, we draw conclusions for nearest neighbor access method design, and for the task of constructing custom access methods tailored to particular applications.

Finding objects in image databases by grouping

Retrieving images from very large collections, using image content as a key, is becoming an important problem. Finding objects in image databases is a big challenge in the field. The paper describes our approach to object recognition, which is distinguished by: a rich involvement of early visual primitives, including color and texture; hierarchical grouping and learning strategies in the classification process; the ability to deal with rather general objects in uncontrolled configurations and contexts. We illustrate these properties with three case studies: one demonstrating the use of color and texture descriptors; one learning scenery concepts using grouped features; and one demonstrating a possible application domain in detecting naked people in a scene.

Region-Based Image Retrieval

As the world becomes an increasingly networked place, effective access to information grows ever more important. This access can take several forms, including traditional database retrieval of structured information, retrieval in collections of documents, and search in collections of binary objects such as sounds, images, and videos. In the latter case, the key challenge is making sense of the objects: if we want to retrieve images of horses, how can we go about processing each image to assess the probability that it contains a horse? This is not a “toy” problem; real users such as graphic designers, editors looking for newspaper photos, students writing reports, and biologists looking for plant or animal specimens need to find images of particular objects.

Window design for signal-dependent spectrogram using optimal kernel techniques

Time-frequency distributions (TFDs) have proven useful in a wide variety of nonstationary signal processing applications. While sophisticated optimal bilinear TFDs have been developed to extract the maximum possible time- frequency information from signals, certain applications dictate simpler linear, running-FFT processing techniques. In this paper, we propose a signal-dependent short-time Fourier transform/spectrogram that enjoys many of the advantages of optimal bilinear TFDs yet retains the simplicity and efficiency of running-FFT processing. In addition, we extend the optimal kernel design problem to linear spaces of signals.