Existing computer programs designed to perform visual recognition of objects suffer from a basic weakness: the inability to spotlight regions in the image that potentially correspond to objects of interest. The brain's mechanisms of visual attention, elucidated by psychophysicists and neurophysiologists, may suggest a solution to the computer's problem of object recognition.

Artifacts made by humans, such as items of furniture and houses, exhibit an enormous amount of variability in shape. In this paper, we concentrate on models of the shapes of objects that are made up of fixed collections of sub-parts whose dimensions and spatial arrangement exhibit variation. Our goals are: to learn these models from data and to use them for recognition. Our emphasis is on learning and recognition from three-dimensional data, to test the basic shape-modeling methodology. In this paper we also demonstrate how to use models learned in three dimensions for recognition of two-dimensional sketches of objects. / Singapore-MIT Alliance (SMA)

No description available.

Thesis (M.S.)--Rochester Institute of Technology, 1988. / Includes appendixes. Includes bibliographical references (leaves 86-87).

Thesis (M.S.)--Rochester Institute of Technology, 1993. / Typescript. Includes bibliographical references (leaves 138-139).

Thesis (Ph.D.(Electrical, Electronic and Computer Engineering))--University of Pretoria, 2008. / Includes bibliographical references (leaves 97-104)

Thesis--(Ph. D.)--University of Wisconsin--Madison, 1971. / Typescript. Vita. eContent provider-neutral record in process. Description based on print version record. Includes bibliography.

Localized operators, like Gabor wavelets and difference-of-Gaussian filters, are considered to be useful tools for image representation. This is due to their ability to form a ‘sparse code’ that can serve as a basis set for high-fidelity reconstruction of natural images. However, for many visual tasks, the more appropriate criterion of representational efficacy is ‘recognition’, rather than ‘reconstruction’. It is unclear whether simple local features provide the stability necessary to subserve robust recognition of complex objects. In this paper, we search the space of two-lobed differential operators for those that constitute a good representational code under recognition/discrimination criteria. We find that a novel operator, which we call the ‘dissociated dipole’ displays useful properties in this regard. We describe simple computational experiments to assess the merits of such dipoles relative to the more traditional local operators. The results suggest that non-local operators constitute a vocabulary that is stable across a range of image transformations.

A real-time character recognition system was developed to recognize upper case handprinted characters in a real-time small machine environment. The recognition system consists of two major components: namely, a data acquisition system and a pattern recognition system. The data acquisition system was designed and implemented to allow the real world data flow into the computer from a COMPUTER writing tablet in real time. The pattern recognition system was also designed and implemented to yield a decision on the input character in real time (user time). A curve optimization technique originally devised by Reumann and Witkam was modified to extract only the significant data that describes a character. Computations were minimized through mathematical simplifications, hardware-software trade-off, and special programming techniques at the machine level. In addition, the preprocessor operated concurrently with the data acquisition routine to reduce data storage requirements as well as to-provide fast response to handprinted inputs. A non-uniform quantization plane was proposed and implemented to discriminate pen directions. Stroke patterns of a character were recognized using a syntactic approach. Finally, recognized stroke patterns within a character were classified as one of the known pattern classes by two classification methods: dictionary look-up and a modified nearest neighbor rule, both guided by special geometric measurements on some character pairs. Character patterns were defined in the dictionary such that no user training or personalized dictionary is required for future use. A test was conducted using the ACM proposed upper case handprinted character set and a recognition rate of 98.3% was obtained from over 2300 characters of sizes varying from 1/4 inch tall to 3/4 inch tall from 10 people. It Is observed that the approach taken in this thesis can also be applied to recognized handprinted patterns other than the standard one proposed by the ACM. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

This thesis presents techniques for locating landmarks in images of human faces. A modified Active Shape Model (ASM [21]) is introduced that uses a form of SIFT descriptors [68]. Multivariate Adaptive Regression Splines (MARS [40]) are used to efficiently match descriptors around landmarks. This modified ASM is fast and performs well on frontal faces. The model is then extended to also handle non-frontal faces. This is done by first estimating the face's pose, rotating the face upright, then applying one of three ASM submodels specialized for frontal, left, or right three-quarter views. The multiview model is shown to be effective on a variety of datasets.