Showing papers on "Scale-invariant feature transform published in 2002"

A Hough transform-based method for fast detection of fixed period sinusoidal curves in images

Abstract: The paper deals with research on detecting a kind of fixed period sinusoidal curve in images A novel method based on the Hough transform is proposed for the purpose, which proceeds on two stages The first stage locates the baselines of sinusoidal curves using a voting process The second stage determines the amplitude and original phase of the curve corresponding to a baseline position using a 2D Hough transform Experimental results are given for synthetic images with and without noise Compared to the standard Hough transform, this method reduces computer storage and computation significantly, and saves much more time

Implementation of Hough transform and its application in line detection and video motion analysis

Abstract: An implementation of Hough transform is described. According to one aspect, the Hough transform is used for line detection in an image based on a one-dimensional voting array.

Noise robust speech recognition using F0 contour extracted by hough transform.

Abstract: This paper proposes a noise robust speech recognition method using prosodic information. In Japanese, fundamental frequency (F0) contour represents phrase intonation and word accent information. Consequently, it conveys information about prosodic phrase and word boundaries. This paper first proposes a noise robust F0 extraction method using Hough transform, which achieves high extraction rates under various noise environments. Then it proposes a robust speech recognition method using syllable HMMs which model both segmental spectral features and F0 contours. Speaker-independent experiments are conducted using connected digits uttered by 11 male speakers in various kinds of noise and SNR conditions. The recognition accuracy is improved in all noise conditions, and the best absolute improvement of digit accuracy is about 4.7%. This improvement is achieved due to the more precise digit boundary detection by the robust prosodic information.

A concurrent modified algorithm for Generalized Hough Transform

Abstract: This paper presents a novel concurrent algorithm for object detection based on the Hough Transform. The Generalized Hough Transform can detect object contours regardless of scale and orientation, but has a computational complexity of O(N/sup 2/RS), where N, R, and S are the array dimensions for X/Y, rotation, and scale, respectively. The high complexity makes it impossible to perform object detection in real-time. In our work, we propose a modified, concurrent algorithm using a multi-threading technique with manager-worker scheme to obtain a reduced complexity of O(N/sup 2//M) where M is the number of processors. Our new algorithm utilizes multi-threading technology to enhance the computing speed. The algorithm is evaluated from both the perspective of output image quality and performance scalability.

Implementation of Hough transform and its application in video motion analysis

Abstract: An implementation of Hough transform is described. According to one aspect, the Hough transform is used to perform motion analysis of patterns in a video using two-dimensional primary data items and a two-dimensional voting array.

Study of image segmentation based on improved Hough transform

Abstract: Hough transform has been widely used in pattern recognition and computer vision since its discovery,but it cannot detect the endpoints and length of a curve. Furthermore, high memory space is required in traditional Hough transform. In this paper, an improved Hough transform algorithm is proposed. According to the inverse transform of HT,it subdivides image space little by little and excludes the area without lines. With this algorithm,the memory space required is reduced, and high efficiency is available, besides the endpoint and length of a curve can be detected. It has been applied to the vehicle license plate recognition (LPR) system successfully.

Research on Measurement of Wave Speed for Sub-Surface Penetrating Radar

Abstract: The measurement of wave speed is always one of the key problems in the research of subsurface penetraing radar, particularly when the location and image of the targets in the ground are measured. On the basis of the signal features of subsurface penetrating radar, extraction of edge contours which uses zero-crossing points and others traditional methods are discarded. The method of extremum is given for extraction of edge contours. The edge contours of the targets can be extracted more accurately using the method of extremum and the new method generates less data for Hough transform. And according to the analysis of Hough transform and the deficiency of traditional Hough transform, an improved Hough transform is proposed in which a FIR low-pass filter is added to improve the accuracy of Hough transform. The improved Hough transform has better performance than the traditional Hough transform by the evaluation for the practical data.

A fuzzy hough transform approach to shape description

Abstract: To use the Hough transform to detect shapes we need to accumulate votes for the edge passing a specific bin. Most existing Hough transform techniques use a sharp (crisp) cutoff to determine whether the bin has received a vote or not. This results in considerable errors. In this paper, we propose a new line Hough transform (LHT) using evidence accumulation and fuzzy aggregation function. The resulting voting process is dependent on the distance ρ from the grid centers. This effectively handles uncertainty in the accumulation process and achieves a better performance. To show the effectiveness this approach, we present our experimental results for a set of 2D parametric and 3D nonparametric objects.

Robust Hough transform LMedS HT and its application to visual inspection

Abstract: In this paper, we propose a new Hough transform algorithm, Least Median of Squares (LMedS) Hough transform, which uses the measure of the least median of squares as the basis to estimate lines. This means that LMedS Hough transform can provide a new measure for finding lines as an alternative to the majority standard of the ordinary Hough transform and, therefore, that LMedS Hough transform can detect lines in the same way as LMedS line fitting procedure. In addition to this, because this algorithm is constructed on the Hough transform paradigm, the basic properties of Hough transform such as noise robustness, multi-line detection and global line detection are inherited in LMedS Hough transform algorithm.