A procedure inside the context of artificial vision that estimates the weight in bovine livestock was developed and designed. The input data to the system are images obtained by a videotape camera to color. These images are digitized and preprocessing using filters of elimination of noise; later are implemented and evaluated different segmentation methods that allow to obtain the contour of the animal in semiautomatic form. This process consists of an automatic binarization in its initial stage and an eventual manual adjustment. Later the characteristics that have a strong correlation with the weight of the animal are extracted. These are extracted in a form that is independent of the orientation of the object inside the image. Such characteristics include: the superior area, perimeter, wide of abdomen, wide of haunch, wide of scapula. The estimate of the weight of the animal is made by means of a neural network type feedforward whose inputs feed with the extracted characteristics of the image. Finally the system is evaluated in the number and type of characteristics kept in mind and in the structure of the neural network utilized.
A method in order to carry out the verification of handwritten signatures is described. The method keeps in mind global features and local features that encode the shape and the dynamics of the signatures. Signatures are recorded with a digital tablet that can read the position and pressure of the pen. Input patterns are considered time and space dependent. Before extracting the information of the static features such as total length or height/width ratio, and the dynamic features such as speed or acceleration, the signature is normalized for position, size and orientation using its Fourier Descriptors. The comparison stage is carried out for algorithms of neurals networks. For each one of the sets of features a special two stage Perceptron OCON (one-class-one-network) classification structure has been implemented. In the first stage networks multilayer perceptron with few neurons are used. The classifier combines the decision results of the neural networks and the Euclidean distance obtained using the two feature sets. The results of the first-stage classifier feed a second-stage radial basis function (RBF) neural network structure, which makes the final decision. The entire system was extensively tested, 160 neurals networks has been implemented.
A method for the verification of handwritten signatures based on the three-dimensional correlation is proposed. The input images to the algorithms of 3D correlation are those range images obtained from the 3D information formed by the signature path. An optic system based on the laser triangulation method, allows the extracting of 3D information. The method takes into account local characteristics that set shape and pressure of signatures together, discrimining characteristics for verification that have been traditionally used separately.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.