Abstract: A pattern searching method using neural networks and correlation. This method combines the quickness and adaptiveness of neural networks with the accuracy of the mathematical correlation approach. Images are divided into small sub-images which are presented to the trained neural network. Sub-images that may contain the pattern or partial pattern are selected by the neural network. The neural network also provides the approximate location of the pattern, therefore the selected sub-images can be adjusted to contain the complete pattern. Desired patterns can be located by measuring the new sub-images' correlation values against the reference models in a small area. Experiments show that this superior method is able to find the desired patterns. Moreover, this method is much faster than traditional pattern searching methods which use only correlation.

Abstract: A user interface provides a uniform control scheme among different devices in a personal communications routing system. The system allows devices such as a personal computer, personal information communicator, pager, fax machine and base station to be operated with a standard twelve key telephone keypad controller for the purpose of accessing electronic messages such as e-mail, voice-mail, fax transmissions, etc. Handheld communicators are docked with the devices to provide wireless transmission between a base station and numerous devices in a multi-user environment. The base station receives electronic messages for the users in, e.g., an office and routes the messages to a device as the user instructs. The user is able to perform operations on the messages such as reading, responding, deleting, etc., by using the keypad.

Abstract: A pattern searching method using neural networks and correlation. This method combines the quickness and adaptiveness of neural networks with the accuracy of the mathematical correlation approach. Images are divided into small sub-images which are presented to the trained neural network. Sub-images that may contain the pattern or partial pattern are selected by the neural network. The neural network also provides the approximate location of the pattern, therefore the selected sub-images can be adjusted to contain the complete pattern. Desired patterns can be located by measuring the new sub-images' correlation values against the reference models in a small area. Experiments show that this superior method is able to find the desired patterns. Moreover, this method is much faster than traditional pattern searching methods which use only correlation.

Abstract: A computer system for use on either a horizontal surface or a vertical wall. The system includes a computer having a base. The system further includes a viewing device for displaying information from the computer. The device is rotatably mounted to the computer to enable rotation of the computer and the device relative to each other. The system further includes a keyboard having keys for inputting commands to the computer. The keyboard is rotatably mounted to the computer to enable rotation of the keyboard relative to the computer. Upon placement of the base on the desk top or vertical wall, the device is rotatable to a substantially vertical position suitable for viewing the device. Further, when the base is positioned on the horizontal surface, the keyboard may be rotated to a substantially horizontal orientation. In addition, when the base is positioned on the vertical wall, the keyboard may be rotated to a substantially vertical orientation suitable for accessing the keyboard.

Abstract: A method of and apparatus for high speed, high resolution, time-shift two-step analog-to-digital conversion (ADC) employing only one ADC module for both coarse and fine signal conversions. The same ADC module is used first for the coarse signal conversion with its output signal stored until the ADC completes the slower fine signal conversion to generate the conversion error for the subsequent compensation process. A digital signal is then generated after the two signals are processed by using a digital signal processing to compensate for conversion error.