Showing papers on "Canny edge detector published in 1994"

Optimal edge detection using expansion matching and restoration

Abstract: Discusses the application of a newly developed expansion matching method for edge detection. Expansion matching optimizes a novel matching criterion called the discriminative signal-to-noise ratio (DSNR) and has been shown to robustly recognize templates under conditions of noise, severe occlusion and superposition. The DSNR criterion is better suited to evaluate matching in practical conditions than the traditional SNR since it considers as "noise" even the off-center response of the filter to the template itself. We introduce a family of optimal DSNR edge detectors based on the expansion filter for several edge models. For step edges, the optimal DSNR step expansion filter (SEF) is compared with the widely used Canny edge detector (CED). Experimental comparisons show that our edge detector yields better performance than the CED in terms of DSNR even under very adverse noise conditions. As for boundary detection, the SEF consistently yields higher figures of merit than the CED on a synthetic binary image over a wide range of noise levels. Results also show that the design parameters of size or width of the SEF are less critical than the CED variance. This means that a single scale of the SEF spans a larger range of input noise than a single scale of the CED. Experiments on a noisy image reveal that the SEF yields less noisy edge elements and preserves structural details more accurately. On the other hand, the CED output has better suppression of multiple responses than the corresponding SEF output. >

Detection of composite edges

Abstract: The paper presents a new parametric model-based approach to high-precision composite edge detection using orthogonal Zernike moment-based operators. It deals with two types of composite edges: (a) generalized step and (b) pulse/staircase edges. A 2-D generalized step edge is modeled in terms of five parameters: two gradients on two sides of the edge, the distance from the center of the candidate pixel, the orientation of the edge and the step size at the location of the edge. A 2-D pulse/staircase edge is modeled in terms of two steps located at two positions within the mask, and the edge orientation. A pulse edge is formed if the steps are of opposite polarities whereas a staircase edge results from two steps having the same polarity. Two complex and two real Zernike moment-based masks are designed to determine parameters of both the 2-D edge models. For a given edge model, estimated parameter values at a point are used to detect the presence or absence of that type of edge. Extensive noise analysis is performed to demonstrate the robustness of the proposed operators. Experimental results with intensity and range images are included to demonstrate the efficacy of the proposed edge detection technique as well as to compare its performance with the geometric moment-based step edge detection technique and Canny's (1986) edge detector. >

Histogram-based morphological edge detector

Abstract: Presents a new edge detector for automatic extraction of oceanographic (mesoscale) features present in infrared (IR) images obtained from the Advanced Very High Resolution Radiometer (AVHRR). Conventional edge detectors are very sensitive to edge fine structure, which makes it difficult to distinguish the weak gradients that are useful in this application from noise. Mathematical morphology has been used in the past to develop efficient and statistically robust edge detectors. Image analysis techniques use the histogram for operations such as thresholding and edge extraction in a local neighborhood in the image. An efficient computational framework is discussed for extraction of mesoscale features present in IR images. The technique presented in the present article, called the Histogram-Based Morphological Edge detector (HMED), extracts all the weak gradients, yet retains the edge sharpness in the image. A new morphological operation defined in the domain of the histogram of an image is also presented. An interesting experimental result was found by applying the HMED technique to oceanographic data in which certain features are known to have edge gradients of varying strength. >

Edge detection and classification using Mallat's wavelet

Abstract: We have developed an improved edge detector and classifier for grey level images using multiresolution wavelet-based analysis, particularly the wavelet introduced by Mallat (see IEEE Trans. on Patt. Anal. and Machine Intell., vol.14, no.7, p.710, 1992), specifically designed for edge detection. The edge detection algorithm has been designed based on a top-down maxima searching criterion, giving the best edge position to that obtained in the lowest scale. We have also been able to classify four different edge profiles: step, ramp, pulse and stair. The classification has been made training a neural network with the coefficients' evolution across scales at the edge localization. The results obtained with a synthetic 256/spl times/256 grey level image with four shapes, each one having a different edge profile have been presented. A perfect segmentation of the four objects is reached. >

Hardware-software codesign of multidimensional programs

Abstract: Presents a method for parametrised partitioning of multidimensional programs for acceleration using a hardware coprocessor. The method involves a divide-and-conquer structure, with the "divide" and "merge" phases carried out by a general-purpose processor, while the "conquer" phase is handled by application-specific hardware. The partitioning strategy has been captured in a simple functional language, and we have automated the production of partitioned programs in this language. Our approach has been tested on an FPGA-based system using a number of computer vision algorithms, including the Canny edge detector, and the performance is compared against executing the programs on the PC host. >

Better geometric measurements based on photometric information

Abstract: We propose new methods for estimating properties of analog objects in properly sampled multi-dimensional grey-scale images. The finite aperture of lenses ensures band limitation of the analog image and allows sampling. Many existing measurement procedures work on a binary object obtained by edge detection and thresholding. The ragged binary edge is disturbed by aliasing which cannot be repaired by smoothing. To solve this problem we propose methods that work directly on the grey-scale image. The grey-scale image contains accurate photometric information. Our new methods yield errors that are generally an order of magnitude better than the traditional binary ones. For applications where a smooth, constant edge height is a prerequisite we introduce erf-clipping. Erf-clipping is a point operation that shapes a linear edge region into a scaled error function. In contrast to thresholding it requires very mild oversampling. >

A zero-crossing-based optimal three-dimensional edge detector

Abstract: Three-dimensional (3D) image processing and interpretation is very important in many medical and industrial applications. Detection of 3D boundaries is an essential step in most of the 3D image analysis tasks. In this paper a new computational approach to 3D edge detection is proposed. Optimality criteria such as signal-to-noise ratio, localization, and spurious response for zero-crossing-based, rotationally invariant 3D step edge detectors are derived. An optimal 3D step edge detector is obtained by optimizing a penalty function which combines all the three criteria. The closed form solution to the optimization problem yields the optimal detector. The detector is the Laplacian of a rotationally invariant function, which has a finite spatial support. The behavior of the proposed detector is theoretically analyzed and compared to that of the 3D Laplacian of Gaussian detector. Experimental results with some synthetic and real images are presented.

Robust consensus based edge detection

Abstract: We present a new robust algorithm for edge detection. The algorithm detects both roof and step type edges. A pixel is declared as an edge pixel if there is a consensus between different processes that try to determine if the pixel lies on a discontinuity. We use robust estimation methods to estimate local fits to windows in the pixel′s neighborhood and accumulate votes from each fit. The use of robust estimators enables us to transform any window possibly containing a discontinuity to a binary window containing a step edge in the location of the discontinuity. We then employ conventional methods to detect this step edge. We show experimental results on simulated edges and synthetic images with varying Gaussian and random noise levels and analyze the probability of detection. The algorithm is also applied to several real intensity and range images and it is shown to perform well. A comparison with the Canny edge detector is given when applicable.

A step edge detector algorithm based on symbolic analysis

Abstract: This paper analyzes the noise error, in both the module and the argument in the calculation of the gradient vector of the images illumination function. The results show that the behavior of the argument is more robust than that of the module. A proposal is made for a symbolic analysis of the argument of the gradient vector to detect the contour of the objects in the image. This shows that it is possible to use smaller windows to calculate the gradient vector without infringing the contradiction proposed by Marr and Hildreth (1980) and Canny (1983). Finally, a description is given of an edge detection algorithm where their most important characteristics are: a) it introduces a symbolic analysis of the argument of the gradient vector to detect edges; and b) It uses smaller window to approximate the value of the gradient vector allowing to locate the edge with precision.

Frame of wavelets for edge detection

Abstract: We present in the following work, a multiscale edge detection algorithm whose aim is to detect edges of any slope. Our work is based on a generalization of the Canny-Deriche filter, modelized by a more realistic edge than the traditional step shape edge. The filter impulse response is used to generate a frame of wavelets. For the merging of the wavelet coefficients, we use a geometrical classifier developed in our laboratory. The segmentation system thus set up and after the training phase does not require any adjustment nor parameter. The main original property of this algorithm is that it leads to a binary edge image without any threshold setting.© (1994) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Training edge detecting fuzzy neural networks with model-based examples

Abstract: A model-based method for training feedforward, backpropagation neural-like networks to produce edge images from data such as forward looking infrared and gray tone pictures is presented. The authors' approach is to train the network on a very small basis set of binary-valued window vectors which are first scored using the Sobel edge operator. Sobel scores are then used to select training vectors that have either crisp or fuzzy edge labels. This training scheme is independent of all real images. The method proposed is illustrated by comparing FF/BP edge images with those produced by the Sobel and Canny edge operators. >

A novel edge detector based on nonlinear local operations

Abstract: A new type of edge detector is presented that uses spatial filtering of the locally thresholded image. The process is realized in two main stages. In the first one an appropriate threshold is used and the multilevel image is transformed to binary one. The second stage verifies the existence of an edge point by correlating the binary valued image points with certain binary valued spatial patterns. False edge points generated by noise are removed by means of another complementary filtering path based on the edge step size. >

A two-dimensional edge detection scheme for general visual processing

Abstract: This paper presents a new two-dimensional (2D) edge detection scheme for general visual processing. The scheme constructs a 2D edge detection functional under the guidance of the Laplacian of Gaussian (LOG) zero crossing contours to detect edges. The detection functional is optimal in terms of signal-to-noise ratio (SNR) and edge localization accuracy (ELA) for detecting edges in 2D images; it also preserves the nice scaling property that is held uniquely by the LOG operator in scale space. The scheme also provides: (1) an edge regularization procedure; (2) an adaptive edge thresholding procedure; and (3) a scale space combining procedure. Experimental results on real images are given in the paper.

A low-cost approach to real-time morphological edge detection

Abstract: Real-time edge-based image detection is an important task in many image analysis operations. Morphological-based edge detection operators have been shown to be effective as well as being efficiently implementable in many image processing applications. In this paper, the concept behind the development of various morphological edge detection operators is briefly described. A new edge detector, based on combining separable median filtering and morphological operations, is introduced. The performance of the proposed edge detector is compared with some other edge detectors. A low-cost methodology for implementing the morphological edge detectors is presented. >

Edge detection by multiscale merging

Abstract: Presents multiscale edge detection on real world images. These images are often 3D and the lens field depths used cause a blur of the edges to be detected. The work is based on a generalization of the Canny-Deriche filter, modeled by a more realistic edge than the traditional ramp form edge. This filter can be used to generate a frame of wavelet. For the merging of the decomposition data, the authors use a geometrical classifier developed in their laboratory. The segmentation system thus set up and after the training phase does not require any adjustment nor parameter. >

Edge detection and curve enhancement using the facet model and parametrized relaxation labelling

Abstract: We present a method for detecting and labeling the edge structures in digital grey-scale images in two distinct stages: 1) a variant of the cubic facet model detects location, orientation and curvature of the putative edge points; and 2) a relaxation labeling network reinforces meaningful edge structures and suppresses noisy edges. Each node label of this network is a 3D vector parametrizing the orientation and curvature of the corresponding edge point. A hysteresis step in the relaxation process maximizes connected contours. For certain images, prefiltering by adaptive smoothing improves robustness against noise and spatial blurring.

A new edge detection method through template matching

Abstract: This paper reviews some gradient edge detection methods and proposes a new detector — the template matching edge detector (TMED). This detector utilizes the concepts of pattern analysis and the template matching of 3×3 masks. A set of performance criteria was used to evaluate the gradient edge detectors as well as the template matching edge detector. The results indicate that the new method is superior to the other gradient edge detectors. In addition, the template matching edge detector has also demonstrated good performance on noisy images. It can obtain very precise edge detection of single pixel width.

Shape feature extraction from object corners

Abstract: A method to extract shape features based on corners is described. Corners contain most of the shape information. Extraction of shape features which are invariant to scaling, rotation and translation is an important problem in computer vision and automatic target recognition systems. A Canny (1986) edge detector which is capable of producing single pixel wide edges is used for obtaining the contour from an image. Using this closed contour as input, the arch height function is computed at each point. The local maxima's correspond to the corner points in the shape. A set of efficient one dimensional moments which are invariant under rotation, translation and scale change is computed. These are the corresponding shape features. Classification is achieved by comparing the extracted features with the shape feature library. In order to validate the concept the following experiments were performed. Ten dissimilar aircrafts and ten similar aircrafts were used as inputs. Contour based moments performed better than the geometric moments in both the data sets. Rotation invariance of two very similar aircrafts showed that contour based moments performed better. The procedure described provides an elegant approach for extracting shape features. These features can also be used as inputs for training and recognizing shapes using neural networks. >

Contribution of Canny-Deriche filter and artificial neural networks to image segmentation

Abstract: In this paper, an image segmentation method based on an edge detection view is presented. This method uses the contribution of two approaches: the optimal edge detector proposed by J. Canny, and then extended to the optimal recursive filter by R. Deriche; and the artificial neural network approach. Combining these two methods, the hysteresis stage needed in the technique developed by Deriche is avoided without damaging the segmentation result. Therefore, thresholds required in hysteresis phase and, which are usually quite difficult to find are no more needed. Experimental results show the validity of this method. >

Use of M-band wavelet transform for multidirectional and multiscale edge detection

Abstract: The problem of directional and multiscale edge detection is considered. Orthogonal and linear-phase M-band wavelet transform is used to decompose the image into channels which correspond to different directions and resolution levels. "Meaningful" combinations of the synthesis filter outputs at the selected channels yield edge maps with desired properties. In addition, with the application of energy operators at the analysis stage, it is possible to obtain a reduction in unwanted zero-crossings. Final edge maps of images are obtained through simple combinations of directional edge maps. >

Optimization approach to edge detection

Abstract: Most methods for finding the edges in a gray-scale image suffer from poor performance in a noisy environment. A novel optimization solution to this problem is explored. An optimization function, based on the output of a threshold-dependent edge detector, is constructed for modeling an a posteriori probability function for the edge image. The prior distribution for the edge image is formed as a Markov random-field model imbued with properties of edge images obtained in a noise-free environment. The edge image that maximizes the optimization function is generated with the simulated annealing algorithm. This edge-detection method, though computationally intensive, explores the use of higher-order Markov random fields that hold rich potential for varied applications. The performance of this approach at four signal-to-noise ratios is compared with that of standard gradient-based edge detection with real and simulated images.

Comparison between MDA and two edge detectors for SAR image analysis

Abstract: A comparison is performed among three edge detectors operating on Synthetic Aperture Radar (SAR) images: Canny filter, Ratio detector and Multilevel Deterministic Annealing with a speckle noise model. The first method represents classical edge extractors, based on a regularised estimation of the derivatives of the image function. The Ratio Detector is an example of a Constant False Alarm Rate methods used for edge detection in speckle-noise, which rely on a detection criterion independent from the average gray-level of the decision region. The method defined Multilevel Deterministic Annealing is based on a non-linear statistical criterion which takes into account not only local properties of the image function, but also its global characteristics. The three detectors are evaluated on the basis of false-alarm and detection errors computed on the basis of the apriori known output of a synthetic scene. A qualitative evaluation on real images is also provided. It is shown that Multilevel Deterministic Annealing provides better results despite of a higher computational load.© (1994) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Efficient algorithm for diffuse edge detection

Abstract: In many digital images, edges do not have a step-like shape but appear as a ramp shape with a very low slope because of noise and blurring effects. In such cases, the white noise model does not hold and the edges are usually called diffuse edges. This paper describes a new second-order gradient based algorithm for the detection of diffuse edges. In the gradient-based edge detection algorithm, such as the LoG filter, there are three basic operations: smoothing as a preprocessing operation to reduce the noise effects, gradient generation - the main operation, and edge linking as a post-processing operation. We propose to use median filtering instead of Gaussian lowpass filtering to reduce the noise, and then introduce a complementary operator of 2D averaging as a second-order gradient generator, and a simple thresholding algorithm to detect zero-crossings. The edge image is finally constructed from edge thinning and linking operations. Simulation results obtained using the proposed edge detection algorithm in comparison with some other commonly used algorithms are included.© (1994) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.