Optimal edge detector design I: parameter selection and noise effects

This paper comments on the optimality of the Laplacian of a Gaussian edge detection filter which localizes edges through zero crossings in the filtered image. The arguments of both Marr and Hildreth, and Dickey and Shanmugam are reviewed to establish that the filter is optimal in the sense of maximizing output image energy near edge features. This filter's principal advantage over other edge detectors is that its response is user-adjustable through selection of a single parameter, the Gaussian standard deviation. However, no clear method for the selection of this parameter has been provided. The problem is addressed here by applying the filter to two ideal periodic edge models blurred by a Gaussian distribution point-spread function. The observed response to the edge spacing and blur standard deviation is then translated into a filter parameter design procedure. The problems of optimum filter performance in the presence of additive Gaussian noise are then addressed. The problem of selecting the sampled filter's coefficient word size is dealt with in a companion paper.     

Efficient synthesis of gaussian filters by cascaded uniform filters

Gaussian filtering is an important tool in image processing and computer vision. In this paper we discuss the background of Gaussian filtering and look at some methods for implementing it. Consideration of the central limit theorem suggests using a cascade of ``simple' filters as a means of computing Gaussian filters. Among ``simple' filters, uniform-coefficient finite-impulse-response digital filters are especially economical to implement. The idea of cascaded uniform filters has been around for a while [13], [16]. We show that this method is economical to implement, has good filtering characteristics, and is appropriate for hardware implementation. We point out an equivalence to one of Burt's methods [1], [3] under certain circumstances. As an extension, we describe an approach to implementing a Gaussian Pyramid which requires approximately two addition operations per pixel, per level, per dimension. We examine tradeoffs in choosing an algorithm for Gaussian filtering, and finally discuss an implementation.     

Best linear unbiased estimators for properties of digitized straight lines

This paper considers the problem of measuring properties of digitized straight lines from the viewpoint of measurement methodology. The measurement and estimation process is described in detail, revealing the importance of a step called ``characterization' which was not recognized explicitly before. Using this new concept, BLUE (Best Linear Unbiased) estimators are found. These are calculated for various properties of digitized straight lines, and are briefly compared to previous work.     

Selecting mating pairs with linear programming techniques

Mate selection can increase progency merit if overall merit is nonlinear for one or more component traits. An index of expected progeny merit could be calculated for all possible mating pairs, and the set of pairs with the highest progeny mean could be selected. There are serious computational problems for more than a few males and females. To select and mate f, females, and m, males, from n of each, with k0 females per male, would require (nf)(nm)f!/(k0!)m evaluations. Linear programming algorithms can determine the optimal strategy efficiently by considering only a subset of these possibilities. Let pi ij be the index of progency merit of the ith sire mated to the jth dam and Xij be the decision variable for that mating (restricted to 0 or 1). Then the problem of selecting mating pairs can be stated as: maximize sigma i sigma j pi ij Xij, subject to sigma i Xij less than or equal to 1, sigma j Xij less than or equal to k0, sigma i sigma j Xij = f, and Xij = 0 or 1. By including an artificial sire and an artificial dam and choosing appropriate merit values for the artificial matings, this problem can be solved by efficient "transportation" algorithms. These algorithms could be used to develop rational mating packages for dairy artificial insemination studs provided that an objective evaluation of progeny merit could be formulated, provided that merit is not simply additively inherited.     

Determination of furazolidone residues in eggs by HPLC followed by confirmation with a diode-array UV/Vis detector

A sensitive HPLC method for the determination of furazolidone residues in eggs (10-1,000 micrograms/kg) is described. Recovery is about 86%. With the aid of a UV/Vis Diode-Array detector confirmation up to the 15-ppb level was possible. In order to test this method with "real" samples, three laying hens received 30 mg each of furazolidone in feed (single dose). The eggs were collected for five days. After five days traces of furazolidone (5 micrograms/kg) could still be detected.     

Practical observations on the methodology for determining 3-methyl-l-histidine using fluorescamine derivatives

The method for the detection and possible quantification of the unusual amino acid 3-methyl-l-histidine has been refined and elucidated. This is an essential prerequisite for the determination of its value as a robust unequivocal index of lean meat protein.     

Tenderness of major muscles from three breed-types of cattle at different times-on-feed

Seventy-five steers (9 to 12 months of age) of Angus (n = 25), Brahman (n = 25) and Brahman × Angus (n = 25) breed-types of known history were fed a high-energy diet and five steers from each breed-type were slaughtered after 0, 56, 112, 168 and 224 days on test. At seven days post mortem, the left side of each carcass was fabricated and eight major primals or subprimals were obtained. Steaks were removed from the shoulder clod, strip loin, tenderloin, top sirloin, knuckle, top round, bottom round and eye of round for Warner-Bratzler shear (WBS) determinations and from the strip loin and bottom round for sensory panel evaluations. Steaks from Angus and Brahman × Angus steers were generally more tender than steaks from Brahman steers. Tenderness of steaks from steers of the three breed-types responded to time-on-feed differently: (a) steaks from Brahman steers improved (P < 0·05) in tenderness with increasing time-on-feed for all muscles except the M. psoas major (tenderloin); (b) with the exception of the M. longissimus muscle (strip loin), the tenderness of steaks from Angus steers did not change (P > 0·05) as time-on-feed increased and (c) for the Brahman × Angus steers, five of the ten muscles studied improved (P < 0·05) in tenderness with increasing time-on-feed. 'Acceptable' tenderness (WBS values of less than 4·5 kg) for most of the muscles was achieved at feeding periods from 112 to 224 days for Brahman, from 0 to 56 days for Angus and from 56 to 168 days for Brahman × Angus.