this paper presents a classification method based on the Kohonen network and its modifications. the research results are presented in this study. the value range was chosen to optimise the net parameters and the quali...
详细信息
this paper presents a classification method based on the Kohonen network and its modifications. the research results are presented in this study. the value range was chosen to optimise the net parameters and the quality classification of machine element input patterns. Parallel, 2-layers net which allows joining of both geometrical technological features was used in the application. Input patters were written in the form a raster grid, in which every raster has a defined number code. Results shows that nets analysed could be utilised in the area of classification of machine elements. (C) 1998 Elsevier Science S.A. All rights reserved.
Handwritten character recognition has been an active area of research. However, because of the recent advancements in mobile devices with limited amount of memory and computational power, efficient and simple algorith...
详细信息
ISBN:
(纸本)9781467358149;9781467358125
Handwritten character recognition has been an active area of research. However, because of the recent advancements in mobile devices with limited amount of memory and computational power, efficient and simple algorithms for both online and offline character recognition have become more appealing. In this work, an efficient character recognition systems is proposed using LDA Analysis followed by a Bayesian discriminator function based on the Mahalonobis distance. Since LDA is tailored for Gaussian distributed data and the samples dimensionality is high, a couple of preprocessing steps have been applied to reduce dimensionality and cluster the data into semi-Gaussian subclasses. In the first step, affine transformations are applied to the training samples in order to make the scheme robust against distortion. Scaling and Rotation are among those popular distortions which have been considered in this work. Inactive pixels are cut off using a simple algorithm in the next step. then, principal component analysis (PCA) and k-means clustering are applied. the results from preprocessing showed a great potential in dimensionality reduction using transformations that can preserve useful information. Numerical results on the MNIST dataset reached 3% error rate which is lower than the other linear approaches. the proposed linear techniques are discussed in a way that make it easier to have a much clearer understanding of the method and why it works compared to the other classification methods.
In order to measure and recognize shaft orbits a fault diagnosis system based on LabVIEW is developed. Integrating data acquisition, signal analysis and processing, information recognition and fault diagnosis, this sy...
详细信息
ISBN:
(纸本)9780769549323
In order to measure and recognize shaft orbits a fault diagnosis system based on LabVIEW is developed. Integrating data acquisition, signal analysis and processing, information recognition and fault diagnosis, this system can automatically measure, analyze, identify the shaft orbits and get fault results. A patternrecognition method on the basis of the invariant moment algorithm is studied with LabVIEW and used to the shaft orbit recognition. the experimental results show that the shaft orbit of a rotor can be identified through this method and the shape is close to the ellipse. So the main fault of the rotor is imbalance. the fault result is verified by spectrum analysis. this study is helpful to develop the online fault diagnosis system based on LabVIEW.
Functional near-infrared spectroscopy (fNIRS) is a non-invasive, low-cost method used to study the brain's blood flow pattern. Such patterns can enable us to classify performed by a subject. In recent research, mo...
详细信息
ISBN:
(纸本)9781728180229
Functional near-infrared spectroscopy (fNIRS) is a non-invasive, low-cost method used to study the brain's blood flow pattern. Such patterns can enable us to classify performed by a subject. In recent research, most classification systems use traditional machine learning algorithms for the classification of tasks. these methods, which are easier to implement, usually suffer from low accuracy. Further, a complex pre-processing phase is required for data preparation before implementing traditional machine learning methods. the proposed system uses a Bi-Directional LSTM based deep learning architecture for task classification, including mental arithmetic, motor imagery, and idle state using fNIRS data. Further, this system will require less pre-processing than the traditional approach, saving time and computational resources while obtaining an accuracy of 81.48%, which is considerably higher than the accuracy obtained using conventional machine learning algorithms for the same data set.
the proceedings contains 170 papers. Topics discussed include image processing, image coding, labelling and classification, medical applications, motion, stereo and three dimensional, image analysis, image interpretat...
详细信息
the proceedings contains 170 papers. Topics discussed include image processing, image coding, labelling and classification, medical applications, motion, stereo and three dimensional, image analysis, image interpretation, image coding and communications, shape description and recognition, image processing applications, computer architectures, image segmentation, neural networks, industrial inspection, filtering and morphology, texture and color, transport, security and remote sensing.
the dual irregular pyramid combines the advantage of adaptivity with a limited computational complexity of neighborhood operations. the levels of the pyramid, dual graphs, are defined only if (hey are planar. We prove...
详细信息
the proceedings contain 117 papers. the special focus in this conference is on Computer Analysis of Images and patterns. the topics include: Performance characterization in computer vision;low-level computational mono...
ISBN:
(纸本)9783540572336
the proceedings contain 117 papers. the special focus in this conference is on Computer Analysis of Images and patterns. the topics include: Performance characterization in computer vision;low-level computational mono and stereo vision;the dual irregular pyramid;analytical results on the quadtree storage-requirements;calculation and estimation of sample statistics of binary images using quadtree data representations;temporal speckle reduction for feature extraction in ultrasound images;noise effects in statistical subpixel patternrecognition;fast iterative reconstruction of band-limited images from non-uniform;anisotropic filtering of mri data based upon image gradient histogram;fast discrete cosine transform approximation for J-PEG image;error diffusion in block truncation coding;on a bound on signal-to-noise ratio in subband coding of gaussian image process;a linear predictor as a regularization function in adaptive image restoration and reconstruction;inversion of convolution by small kernels;a model-based image quantization technique for supervised image recognition;brightness-contrast diffusion and the grouping of missing angles;using eigenvectors of a vector field for deriving a second directional derivative operator for color images;crest lines detection in grey level images;a comparative study of performance for noisy roof edge detection;a hough-like prediction/correction approach for ellipse detection;fourier parameterization provide uniform bounded hough space;circle extraction via least squares and the kalman filter;a multiresolution shape description algorithm;image coding by morphological skeleton transformation;a non-linear shape abstraction technique and minimum-space time-optimal convex hull algorithms.
this study revolves around NAO, a programmable and interactive robot, to do Emotion recognition via Facial Expression, and to give an appropriate response for the identified emotion whilst accumulating images to furth...
详细信息
ISBN:
(数字)9781728167916
ISBN:
(纸本)9781728167916
this study revolves around NAO, a programmable and interactive robot, to do Emotion recognition via Facial Expression, and to give an appropriate response for the identified emotion whilst accumulating images to further enhance the accuracy. To develop this cohesive system, the authors utilized Computer Vision and Haar-Cascade Classifier along withthe feature descriptors, Histogram of Oriented Gradient and Local Binary pattern to train the Support Vector Machine Learning Algorithm. Gathering and increasing data with Asian/Filipino participants whilst being added to the CK+ database in different stages of retraining, optimization, and undertaking cross-fold validation yielded improved results withthe highest average accuracy across the seven emotions of 87.14%. Suggesting that adding images with a specific ethnicity yields a higher accuracy model for a more diverse testing dataset and in the wild image classifications. Withthe machine learning model capable of accurately recognizing the emotion and reinforcing it withthe accumulated local images, NAO equipped with emotion recognition capabilities can better aid and support the individuals in need.
this paper develops and compares two fuzzy logic based and a traditional rule-based patternrecognition system, which perform target recognition with data from a typical range and doppler resolving radar. the paramete...
详细信息
ISBN:
(纸本)0780336461
this paper develops and compares two fuzzy logic based and a traditional rule-based patternrecognition system, which perform target recognition with data from a typical range and doppler resolving radar. the parameters used are target altitude, velocity, range from nearest base, and radar cross section. the systems identify four classes of aircraft: fighter/interceptors, large bombers, rotary craft, and vertical take off and landing (VTOL) combat aircraft. the first fuzzy based technique classifies targets by selecting the aircraft withthe maximum summed amount of membership, giving a classification accuracy of 94% (average). the second approach classifies targets by selecting the aircraft through a max-min fuzzy decision system. this results in a 99% average accurate classification. the traditional rule-based method implements an expert system and correctly classifies 75% (average) of the targets.
暂无评论