This paper describes an algorithm to estimate the text skew angle in a document image. The algorithm utilizes the recursive morphological transforms and yields accurate estimates of text skew angles on a large documen...
Magnetic Ink Character recognition (MICR) technology has widely been used for processing.bank checks. Since the MICR character set is a special type font, and the ink is also readable by human being, optical approach ...
详细信息
Point Distribution Models (PDMs) are statistical models which represent objects whose shape can vary. A useful feature of PDMs is their ability to capture the shape of variable objects within a training set with a sma...
详细信息
Point Distribution Models (PDMs) are statistical models which represent objects whose shape can vary. A useful feature of PDMs is their ability to capture the shape of variable objects within a training set with a small number of shape parameters. This compact and accurate parametrization can be used for the design of efficient classification systems. In this paper we describe a classification system which uses shape parameters. We have tested the system on classifying hand outlines, face outlines and hand gestures;experimental results are presented.
We present a fully automated process to scan the Australian Telecom Yellow Pages and produce a text document consisting only of the business entries, while removing the advertisements, graphics and notes about the Yel...
详细信息
We describe a novel layered coding scheme in which MPEG encoded video data is split into layers according to the class of each coded picture. We investigate the tolerance of each layer to bit error effects. We conclud...
详细信息
We describe a novel layered coding scheme in which MPEG encoded video data is split into layers according to the class of each coded picture. We investigate the tolerance of each layer to bit error effects. We conclude that the separate layers can tolerate different error rates. By applying different levels of FEC to the separate layers it is possible to ensure reliable transmission whilst minimising the overhead due to FEC.
The edge detection and the restoration of images degraded by LSI and additive Gaussian white noise using an adaptive multiresolution Hopfield network was investigated. Results obtained indicate that a technique that u...
详细信息
The edge detection and the restoration of images degraded by LSI and additive Gaussian white noise using an adaptive multiresolution Hopfield network was investigated. Results obtained indicate that a technique that uses the edge estimates produced by an edge detection network from the distorted data is effective in restoring blurred, noisy images, giving results which both look better and have better SNR's than those obtained using a single resolution. Basing from the results obtained, it is concluded that multiresolution Hopfield networks provide a useful addition to the set of techniques for solving ill-posed imageprocessing.problems.
The fractal transform is a recently developed tool for image analysis and patternrecognition. Problems relating to pixel based image resolution are highlighted. The fractal transform is outlined and shown to be a sol...
详细信息
ISBN:
(纸本)0819418455
The fractal transform is a recently developed tool for image analysis and patternrecognition. Problems relating to pixel based image resolution are highlighted. The fractal transform is outlined and shown to be a solution to pixel based image resolution related problems. MatchMaker, a technique based on fractal transform analysis is analyzed and its ability to find objects in complex scenes is confirmed. A new neural network paradigm, called the fractal transform network, is described. Built from neural network elements commonly described in the classic neural network literature, it is shown that this new paradigm can implement the MatchMaker technique and other fractal transform processes as highly parallel networks.
The present work introduces a novel method of interframe coding using displacement vectors generated by the analysis of classified feature motion. This technique is able to compensate for uniform and complex motion in...
详细信息
The present work introduces a novel method of interframe coding using displacement vectors generated by the analysis of classified feature motion. This technique is able to compensate for uniform and complex motion in varying degrees of efficiency. Errors are introduced by way of quantisation and mode-value blocks to lower image resolution. It is shown that the net propagation of errors in a sequence is not evident. With the use of spatial filtering, errors to be removed are reconstructed, hence improving the perceived spatial and temporal quality of the video sequence.
A method based on connected area detection is developed to segment and recognize dimension texts in engineering drawings. First we propose an effective algorithm to j h d all connected areas from druwing image. Then w...
详细信息
Some research institutes are elaborating plans to develop super high definition (SHD) image data beyond the HDTV. The SHD images should have high spatial and temporal resolutions. Picture quality can be enhanced by in...
详细信息
Some research institutes are elaborating plans to develop super high definition (SHD) image data beyond the HDTV. The SHD images should have high spatial and temporal resolutions. Picture quality can be enhanced by incorporating imageprocessing.techniques into imaging process. Imaging methods based on this approach can be classified into two main categories;a spatial integration imaging method and a temporal imaging method. Just recently, a spatial integration imaging method that performs well has been presented. In the present study, a temporal integration imaging method that uses a single camera and produces an improved-resolution and/or low SNR image by integrating multiple consecutive image frames taken with the camera is presented. Experimental simulations are conducted to demonstrate that this method works well.
暂无评论