No description available.

No description available.

No description available.

One of the major problems in computer vision involves dealing with uncertain information. Occlusion, dissimilar views, insufficient illumination, insufficient resolution, and degradation give rise to imprecise data. At the same time, incomplete or local knowledge of the scene gives rise to imprecise interpretation rules. Uncertainty arises at different processing levels of computer vision either because of the imprecise data or because of the imprecise interpretation rules. It is natural to build computer vision systems that incorporate uncertainty reasoning. The Dempster-Shafer (D-S) theory of evidence is appealing for coping with uncertainty hierarchically. However, very little work has been done to apply D-S theory to practical vision systems because some important problems are yet to be resolved. / Ph. D.

Holistic scene understanding is one of the major goals in recent research of computer vision. Most popular recognition algorithms focus on semantic understanding and are incapable of providing the global depth information of the scene structure from the 2D projection of the world. Yet it is obvious that recovery of scene surface layout could be used to help many practical 3D-based applications, including 2D-to-3D movie re-production, robotic navigation, view synthesis, etc. Therefore, we identify scene geometric reasoning as the key problem of scene understanding. This PhD work makes a contribution to the reconstruction problem of 3D shape of scenes from monocular images. We propose an approach to recognise and reconstruct the geometric structure of the scene from a single image. We have investigated several typical scene geometries and built a few corresponding reference models in a hierarchical order for scene representation. The framework is set up based on the analysis of image statistical features and scene geometric features. Correlation is introduced to theoretically integrate these two types of features. Firstly, an image is categorized into one of the reference geometric models using the spatial pattern classi cation. Then, we estimate the depth pro le of the speci c scene by proposing an algorithm for adaptive automatic scene reconstruction. This algorithm employs speci cally developed reconstruction approaches for di erent geometric models. The theory and algorithms are instantiated in a system for the scene classi cation and visualization. The system is able to fi nd the best fi t model for most of the images from several benchmark datasets. Our experiments show that un-calibrated low-quality monocular images could be e fficiently and realistically reconstructed in simulated 3D space. By our approach, computers could interpret a single still image as its underlying geometry straightforwardly, avoiding usual object occlusion, semantic overlapping and defi ciency problems.

No description available.

Color vision is widespread in the animal kingdom, and describes the ability to discriminate between objects purely based on the wavelengths that they reflect. Experiments across many species have isolated wavelength comparison in the brain as a computation underlying color vision. This comparison takes place in color opponent neurons, which respond with opposite polarity to wavelengths in different parts of the spectrum. In this work, I explore color opponency in the genetically tractable organism Drosophila melanogaster, where these circuits have only just begun to be described. Using two-photon calcium imaging, I measure the spectral tuning of photoreceptors in the fruit fly and identify circuit mechanisms that give rise to opponency. I find two pathways: an insect-specific pathway that compares wavelengths at each point in space, and a horizontal-cell-mediated pathway similar to that found in mammals. The horizontal-cell-mediated pathway enables additional spectral comparisons through lateral inhibition, expanding the range of chromatic encoding in the fly. Together, these two pathways enable efficient decorrelation and dimensionality reduction of photoreceptor signals while retaining maximal chromatic information. This dual mechanism combines motifs of both an insect-specific visual circuit and an evolutionarily convergent circuit architecture, endowing flies with the ability to extract chromatic information at distinct spatial resolutions.

A subcutaneous injection of sodium selenite (30 umol/kg body weight) in 10â day old rats produced a cataract within 72 hours. Lens opacification was preceded by a 15% decrease in ATP content. Lens ATP did not fully recover to control concentrations by 11 days post-injection. A moderate correlation existed between lens weight and total ATP content in control lenses <r=O.509, n=153), but this relationship disappeared with selenite treatment <r=0.023, n=153). There was a significant inhibition by selenite on lens growth from 4 through 11 days after an injection. / Master of Science

Vision systems are widely used in robot workcells for sensory feedback. The resolution of a vision system is usually good enough to locate an object so that it can be grasped, but not good enough to accurately locate an insertion hole. Tactile probes are used to accurately locate objects. However, they require a data base containing the approximate location of an object in order to be used effectively. This thesis presents the development of a robot workcell which utilizes a vision system and tactile probe to identify, locate, and orientate two types of circuit board fixtures. The vision system approximately locates the corner points of each fixture in the robot workcell. The tactile system uses the data base created by the vision system to conduct a tactile search for each fixture and to accurately define the coordinates of each corner point. After a fixture is accurately located, a region (sub-coordinate system) is defined about the fixture. The location of each insertion hole within a fixture is defined relative to the region and the robot subsequently inserts the tactile probe into each hole. The vision system developed can define any two dimensional object and can locate the corner points of any straight edged object, whose adjacent sides have an included angle greater than 90 degrees. The tactile system is self calibrating and has a repeatability of 0.009 inches. A probe insertion error analysis was conducted on the system. The average probe insertion error for the system was determined to be 0.0337 inches. In addition, it was determined that probe insertion error increases with the distance between a hole and the origin of its defining region, and that the major source of probe insertion error is the robot language's (AML/E Verion 4.0) inability to accurately define points within a region. / M.S.

The premise that mature visual function depends upon the nature of visual experience during development is based primarily on experiments showing that visual deprivation during a 'critical' period early in life causes abnormalities in visual cortex and an enduring loss of spatial vision (amplyopia). There is, however, little evidence that early visual experience atually enables mature vision. Experments in this thesis provide such evidence. The measurement of optomotor responses daily from eye opening permanently enhances optomotor sensitivity and the perception of visual motion. The plasticity allowing this enhancement is transient and peaks in efficacy before the start of the classical 'critical ' period for ocular dominance plasticity. The enhancement is dependent upon optomotor responses generated by the movement of high spatial frequency visual stimuli, and is mediated by the visual cortex. These studies show that a form of experience-dependent plasticity, distinct from that of the critical period, enables mature motion vision. / viii, 107 leaves : ill. (some col.) ; 28 cm.