The correspondence problem (finding matching regions in images) is a fundamental task in computer vision. While the concept is simple, the complexity of feature detectors and descriptors has increased as they provide more efficient and higher quality correspondences. This complexity is a barrier to developers or system designers who wish to use computer vision correspondence techniques within t...

Simultaneous Localization and Mapping (SLAM) is the problem of building a map of an unknown environment by a mobile robot while at the same time navigating the environment, using the unfinished map. For SLAM, two tasks have to be solved: First reliable feature extraction and data association, second the optimal estimation of poses and features. These two parts are often referred to as SLAM fron...

552 Abstract—In many documents such as admission form, bank cheques, memorandums, letters and application forms machine printed and handwritten characters are mixed. Since the algorithms for recognition of machine-printed texts and handwritten texts are different, it is necessary to distinguish between these two types of texts before giving it to respective OCR systems to process it. This separ...

I present various systems for the recognition of human faces. They consist of three steps: feature extraction, solving the correspondence problem, and the actual comparison with stored faces. Two of them are implemented in the Dynamic Link Architecture and are, therefore, close to biological hardware, the others are more technical in nature but also have some biological plausibility. At the end...

Recognizing a word in a meaningful text involves processes that combine information from many different sources, and poth bottom-up processes (such as feature extraction and letter recognition) and top-down processes (contextual information) are thought to interact when skilled readers recognize words. Two similar experiments investigated word recognition latencies of adult readers who read sin...

Hansson (2001), Rose & Walker (2004), and Walker (2000a, 2000b) have recently proposed that long-distance consonant assimilation is accomplished via segmental correspondence rather than autosegmental linking. The phonology of the feature [anterior] in Chumash supports this idea: linking of the feature [anterior] is forbidden across morpheme boundaries, but long-distance [anterior] harmony is al...

Abstract The bulk-boundary correspondence is an integral feature of topological analysis and the existence boundary or interface modes offers direct insight into structure Bloch wave function. While only topology function has been considered relevant to modes, we demonstrate that another geometric quantity, so-called quantum distance, can also host a bulk-interface correspondence. We consider g...

This paper investigates a new approach to formulate performance indices of biometric system using information theoretic models. The performance indices proposed here (unlike conventionally used FAR, GAR, DET etc.) are scalable in estimating performance of large scale biometric system. This work proposes a framework for identification capacity of a biometric system, along with insights on number...

A novel minutiae-based fingerprint matching algorithm is proposed. A minutiae matching algorithm has to solve two problems: correspondence and similarity computation. For the correspondence problem, we assign each minutia two descriptors: texture-based and minutiae-based descriptors, and use an alignment-based greedy matching algorithm to establish the correspondences between minutiae. For the ...

This paper introduces a fast image mosaicing technique that does not require costly search on image domain (e.g., pixel-to-pixel correspondence search on the image domain) and the iterative optimization (e.g., gradient-based optimization, iterative optimization, and random optimization) of geometric transformation parameter. The proposed technique is organized in a two-step manner. At both step...