In this paper we present a simple and computationally efficient method for document image enhancement. We use Unsharp masking to enhance the edge detail information in the degraded document. This image is then used to...
详细信息
ISBN:
(纸本)9781424418343
In this paper we present a simple and computationally efficient method for document image enhancement. We use Unsharp masking to enhance the edge detail information in the degraded document. This image is then used to adjust the local threshold for each pixel. Proposed method is experimentally compared with Laplacian sign method and the Otsu method. It is shown that the method improves sharpness of the image with nearly half computational time.
In this paper, a new thresholding approach for data denoising is presented. The approach is based minimum noiseless description length (MNDL), a new method for optimum sub-space selection in data representation. By us...
详细信息
In this paper, a new thresholding approach for data denoising is presented. The approach is based minimum noiseless description length (MNDL), a new method for optimum sub-space selection in data representation. By using the observed noisy data, this information theoretic approach provides the optimum threshold that minimizes the description length of the noiseless signal. Comparison of the new method with the existing thresholding methods is provided
Face recognition and verification under varying pose and illumination is still a challenging problem. In this work, a novel approach for pose invariant face recognition based on artificial neural network under similar...
详细信息
Face recognition and verification under varying pose and illumination is still a challenging problem. In this work, a novel approach for pose invariant face recognition based on artificial neural network under similar illumination condition is proposed. The neural network is trained to learn the face features with variation of pose and interpolate the face features for any unknown pose, leading to a good matching with the probe image. The simple implementation by simulation experiment with HOIP data base shows promising result
This paper presents a neural network based approach for vehicle classification. The proposed vehicle classification approach extracts various features from a vehicle image, normalises and classifies them into one of t...
详细信息
This paper presents a neural network based approach for vehicle classification. The proposed vehicle classification approach extracts various features from a vehicle image, normalises and classifies them into one of the known classes. It is based on structural features and a direct solution training method. The preliminary experiments on training and testing of 4 types of vehicles patterns were conducted. The experimental results are very promising and demonstrate the effectiveness and usefulness of the proposed approach
With this work it is intended to do the tracking and analysis of the human motion, more specifically the gait. By using computational vision, it has been acquired the trajectories of defined control points in individu...
详细信息
ISBN:
(纸本)9781424408290
With this work it is intended to do the tracking and analysis of the human motion, more specifically the gait. By using computational vision, it has been acquired the trajectories of defined control points in individuals' body, throughout time and space. These results are to he used afterwards in gait specification of biped robots. Several types of movement and the phases that compose a common system of capture and analysis of movement are referenced. Then, methods used in imageprocessing and a description of existing gait types are detailed. Finally, the implemented software is presented and the results analyzed.
Traditional Fourier descriptor suffers the drawback of being sensitive to translation, rotation and scale transform. This paper presents a new method of shape representation, polygon Fourier descriptor (PFD), which ca...
详细信息
Traditional Fourier descriptor suffers the drawback of being sensitive to translation, rotation and scale transform. This paper presents a new method of shape representation, polygon Fourier descriptor (PFD), which can achieve the invariance on translation, rotation and scale transform naturally and without need of a process of normalization. By applying polygonal approximation to the contour firstly, the contour can be represented more efficiently. Moreover, PFD is robust enough to represent the contour even if some transformations happened to it. Finally, we apply PFD to the recognition of static gestures and obtain satisfying results
The 2005 DARPA Grand Challenge was a 213 Km robot race across the Mojave Desert between Nevada and California. Our team attempted to run the course using only stereo vision. I designed a novel real-time large-image st...
详细信息
The 2005 DARPA Grand Challenge was a 213 Km robot race across the Mojave Desert between Nevada and California. Our team attempted to run the course using only stereo vision. I designed a novel real-time large-image stereo vision system that uses sparse stereo and is biologically inspired. It uses a filter similar to that used in primary visual cortex to locate small line segments. These are then arranged to form a polyline representation of the scene and estimate the depth. The technique was capable of nine frames per second of a megapixel image on a single processor. Accuracy was questionable and the robot did not compete in the race for other reasons
We spend more and more time on travelling. Travelling by metro or tram gives us almost certain time to get home. And what about is travelling by car? Many persons travel hours to go working or to get home. In rush tim...
详细信息
We spend more and more time on travelling. Travelling by metro or tram gives us almost certain time to get home. And what about is travelling by car? Many persons travel hours to go working or to get home. In rush time the main roads and highways are full of cars. Can we simulate drivers' behaviour and their decision making process? This paper tries to find patterns in drivers' behaviour and simulate these processes.
Intelligent techniques of harmonic detection or estimation are nowadays of a great interest in power system applications, their ability to deal with high non-linearities attract researchers to investigate the performa...
详细信息
ISBN:
(纸本)9781424408290
Intelligent techniques of harmonic detection or estimation are nowadays of a great interest in power system applications, their ability to deal with high non-linearities attract researchers to investigate the performance of these methods mainly based on artificial intelligence namely using artificial neural networks. In the literature many harmonic detection or estimation methods were presented, in this paper we focus on a new idea to apply an adaptive linear neuron (ADALINE) and a multi-layer artificial neural net work (M-LANN) [5] to estimate the fundamental component and the total harmonic content of a distorted signal.
In this paper, we propose a flexible fully-multiplicative orthogonal-group based ICA (FlexibleOgICA) algorithm, which can instantaneously separate the mixture of sub-Gaussian and super-Gaussian source signals. It adop...
详细信息
ISBN:
(纸本)9781424407101
In this paper, we propose a flexible fully-multiplicative orthogonal-group based ICA (FlexibleOgICA) algorithm, which can instantaneously separate the mixture of sub-Gaussian and super-Gaussian source signals. It adopts a self-adaptive nonlinear function, which adjusts its parameter to achieve better performance based on the estimation of the kurtosis of super-Gaussian source signals. We also have successfully applied the algorithm to obtain the fetal electrocardiogram (FECG) signal, showing its fast convergence speed and high separation performance.
暂无评论