An artificialneural network approach was evaluated in multispectral imageprocessingapplications, including general land cover classification and land use feature identification. The performance of an artificial neu...
详细信息
An artificialneural network approach was evaluated in multispectral imageprocessingapplications, including general land cover classification and land use feature identification. The performance of an artificialneural network was compared to that of a standard statistical classification technique. The results show that the neural network is more responsive to cultural materials with skewed spectral distributions and sub-pixel spatial resolution (e.g. asphalt roads). The neural network's probabilistic output also enables discrepancy resolution capabilities for the higher level task of land use feature identification. A model was developed to identify roads from the multispectral classifications. The neural network based model performed much better than the model based on the standard statistical classification technique.
The Lister Hill National Center for Biomedical Communications is a Research and Development Division of the National Library of Medicine. One of the Center's current research projects involves the conversion of en...
详细信息
ISBN:
(纸本)0819412015
The Lister Hill National Center for Biomedical Communications is a Research and Development Division of the National Library of Medicine. One of the Center's current research projects involves the conversion of entire journals to bitmapped binary page images. In an effort to reduce operator errors that sometimes occur during document capture, three back error propagation networks were designed to automatically identify journal title based on features in the binary image of the journal's front cover page. For all three network designs, twenty five journal titles were randomly selected from the stored database of image files. Seven cover page images from each title were selected as the training set. For each title, three other cover page images were selected as the test set. Each bitmapped image was initially processed by counting the total number of black pixels in 32-pixel wide rows and columns of the page image. For the first network, these counts were scaled to create 122-element count vectors as the input vectors to a back error propagation network. The network had one output node for each journal classification. Although the network was successful in correctly classifying the 25 journals, the large input vector resulted in a large network and, consequently, a long training period. In an alternative approach, the first thirty-five coefficients of the Fast Fourier Transform of the count vector were used as the input vector to a second network. A third approach was to train a separate network for each journal using the original count vectors as input and with only one output node. The output of the network could be 'yes' (it is this journal) or 'no' (it is not this journal). This final design promises to be most efficient for a system in which journal titles are added or removed as it does not require retraining a large network for each change.
The detection and localization of faces in an image has many applications in various domains: surveillance, TV audience polling, etc. We propose a new method for this task. The main idea of our method is to train a ne...
详细信息
The detection and localization of faces in an image has many applications in various domains: surveillance, TV audience polling, etc. We propose a new method for this task. The main idea of our method is to train a neural network to detect the presence or absence of a face in its input window, and to scan this network over at all possible locations in the image. Because of the nature of the neural network architecture we used, this process can be done very efficiently without requiring to actually recompute the entire network state at each location. The scanning is performed on several versions of the image at various scales, resulting in an efficient, scale independent detector and locator.
Systems for high-precision control of the trajectory to be followed by a robot arm grip need to properly model the interaction among the robot arm joints and to cope with the high-speed and nonlinearities of the arm d...
详细信息
ISBN:
(纸本)0819412007
Systems for high-precision control of the trajectory to be followed by a robot arm grip need to properly model the interaction among the robot arm joints and to cope with the high-speed and nonlinearities of the arm dynamics. To solve this problem the use of a hardware accelerator, which is able to explore parallelism within multivariable self-tuning control algorithms, is proposed. The accelerator works as part of an integrated system which incorporates facilities of computer vision and robot arm trajectory definition. The computer vision sub-system recognizes the position of an object selected to be picked by the robot arm and the trajectory definition sub-system uses a neural network to define the angular position of the joints along the trajectory to be followed by the arm.
Research in the last decade emphasized the potential of designing adaptive pattern recognition classifiers based on algorithms using multi-layered artificialneural nets. The greatest potential in such endeavors was a...
详细信息
ISBN:
(纸本)0819413267
Research in the last decade emphasized the potential of designing adaptive pattern recognition classifiers based on algorithms using multi-layered artificialneural nets. The greatest potential in such endeavors was anticipated to be not only in the adaptivity but also in the high-speed processing through massively parallel VLSI implementation and optical computing. Computational advantages of such algorithms have been demonstrated in a number of papers. neuralnetworks particularly the self-organizing types have been found quite suitable crisp pattern for clustering of unlabeled datasets. The generalization of Kohonen-type learning vector quantization (LVQ) clustering algorithm to fuzzy LVQ clustering algorithm and its equivalence to fuzzy c-means has been clearly demonstrated recently. On the other hand, Carpenter/Grossberg's ART-type self organizing neuralnetworks have been modified to perform fuzzy clustering by a number of researches in the past few years. The performance of such neuro-fuzzy models in clustering unlabeled data patterns is addressed in this paper. A recent development of a new similarity measure and a new learning rule for updating the centroid of the winning cluster in a fuzzy ART-type neural network is also described. The capability of the above neuro-fuzzy model in better partitioning of datasets into clusters of any shape is demonstrated.
artificialneural net models have been studied for many years in the hope of achieving human-like performance in the fields of speech, image recognition and pattern recognition. For high performance and for controllin...
ISBN:
(纸本)0819412813
artificialneural net models have been studied for many years in the hope of achieving human-like performance in the fields of speech, image recognition and pattern recognition. For high performance and for controlling the size of the network, the input information must be preprocessed before being fed into the neural network. In this paper, a probabilistic spectral feature extraction technique (PSFET) with multiview spectral representations and its applications are described. During training and testing, the PSFET allows efficient extraction of useful information in addition to generating an input vector size for best classification performance by the following neural network. Experimental results indicate that the performance of the neural network increases in classification accuracy when PSFET is used at the input. The network also generalizes better.
An overview of ongoing research related to the development of an image data compression algorithm using artificialneuralnetworks (ANNs) is presented. The data compression technique under study uses an ANN to perform...
详细信息
An overview of ongoing research related to the development of an image data compression algorithm using artificialneuralnetworks (ANNs) is presented. The data compression technique under study uses an ANN to perform vector quantization (VQ). A good predictor is one of the essential components of the image compression technique being explored. The performance of the various predictors are compared including an average predictor, a median predictor, a recurrent artificialneural network (RANN) predictor, and a second-order optimal linear predictor. It is shown that, for some cases, a relatively simple recurrent artificialneural network predictor performs close to the second-order optimal linear predictor and better than the average and the median predictors.< >
Regularization is a paradigm for performing image segmentation and edge detection, that can be implemented in a neural network type architecture. Various topics and problems pertaining to the use of regularization for...
详细信息
ISBN:
(纸本)0819408743
Regularization is a paradigm for performing image segmentation and edge detection, that can be implemented in a neural network type architecture. Various topics and problems pertaining to the use of regularization for imageprocessingapplications are discussed. Topics include data fusion, sensor blur, and the operation on partitioned images. A mathematical analysis of the different topics is presented, including a modification of the original regularization energy functional to perform data fusion.
Holographic photothermoplastic disk with two-dimensional Fourier-holograms registration under infrared laser heating of the recording layer is realized. It supplies high information recording density (105 bit/mm2), an...
详细信息
ISBN:
(纸本)0819408743
Holographic photothermoplastic disk with two-dimensional Fourier-holograms registration under infrared laser heating of the recording layer is realized. It supplies high information recording density (105 bit/mm2), and high quality of Fourier-holograms registration ((eta) equals 5 - 7%, contrast 70:1), in addition to the possibility of information rerecording and invariance to the shift of thin phase holograms. Holographic disk based scheme of the optical neural network with outer product implementation is suggested. Results of such neural network modeling are described.
暂无评论