초록 ▼

A method and apparatus for on-line predicting an analytical model of a dynamically varying color reproduction device comprises recursively executing a parameter adjustment algorithm for updating the parameters to the analytical model. Parameter adjustment is computed based upon a measured error sign

A method and apparatus for on-line predicting an analytical model of a dynamically varying color reproduction device comprises recursively executing a parameter adjustment algorithm for updating the parameters to the analytical model. Parameter adjustment is computed based upon a measured error signal detected between a sensed signal detected from actual device output and with an insitu sensor, and the analytical model output. The adjustment algorithm is recursively executed until parameter convergence effects a minimal error measurement, at which time the updated parameters are identified as accurate within the selected parametric model.

대표청구항 ▼

1. A method for online predicting an analytical model for a dynamically varying color reproduction device including a sensor for detecting a signal representative of device color output, wherein color processing of the device is represented by a preselected form of an analytical model comprising ini

1. A method for online predicting an analytical model for a dynamically varying color reproduction device including a sensor for detecting a signal representative of device color output, wherein color processing of the device is represented by a preselected form of an analytical model comprising initially an inaccurate set of parameters, the method including steps of:producing an image with the device in response to an input signal;sensing with the sensor an output signal representative of the image;computing from the model a model output in response to the input signal and comparable to the image output signal;determining an error between the image output signal and the model output;generating an updated set of parameters for the model from a prediction algorithm utilizing the error and applying the updated set to the model; and,recursively executing said steps with subsequent input signals until said error is less than a predetermined value whereby the analytical model including the updated set of parameters comprises an accurate representation of device operation. 2. The method as defined in claim 1 wherein the generating comprises converging the inaccurate set to the updated set with less than ten recursive executions of said steps. 3. The method as defined in claim 1 wherein the computing comprises utilizing either a linear affine model, a partial quadratic model, a quadratic model or a cubic model as the analytical model. 4. The method as defined in claim 3 wherein the determining comprises using a recursive convergent least squares estimation algorithm as the prediction algorithm. 5. The method as defined in claim 1 further comprising storing the analytical model as a diagnostic program in a network system having a plurality of the reproduction devices. 6. The method as defined in claim 1 wherein the color processing is represented by a plurality of analytical models each associated with a corresponding partition of a gamut of the reproduction device, and wherein the recursively executing said steps comprises the on-line predicting for each of the plurality of analytical models. 7. The method as defined in claim 1 wherein producing comprises applying a test target signal as the input signal. 8. The method as defined in claim 7 further including constructing an inverse of the analytical model including the updated set of parameters and using the inverse for generating a LUT or adjustment of the subsequent input signals. 9. A method for on-line predicting an analytical model of a dynamically varying color reproduction device including a sensor for detecting a signal representative of device color output, wherein color processing of the device is represented by a preselected form of an analytical model comprising initially a random set of parameters, the method including steps of:producing an image with the device in response to an input signal;sensing with the sensor an output signal representative of the image;computing from the model a model output in response to the input signal;determining an error between the image output signal and the model output;generating an updated set of parameters for the model from a prediction algorithm utilizing the error and applying the updated set to the model;inverting the model;adjusting a subsequent input signal with the inverted model; and,recursively executing said steps with a plurality of subsequent input signals until said error is less than a predetermined value whereby the analytical model comprises an accurate representation of device operation. 10. The method as defined in claim 9 wherein the producing comprises applying a selected test target signal as the input signal. 11. The method as defined in claim 9 wherein the analytical model is stored as a diagnostic program in a network system comprising a plurality of the devices, the method further comprising identifying devices in the network generating an error greater than a second predetermined value, whereby the devices h aving varying color reproduction due to operational drift are easily identified. 12. The method as-defined in claim 9 wherein the computing comprises utilizing either a linear affine model, a partial quadratic model, a quadratic model or a cubic model as the analytical model. 13. The method as defined in claim 12 wherein the generating comprises utilizing a recursive convergent least square estimation algorithm as the prediction algorithm. 14. The method as defined in claim 9 wherein the color processing is represented by a plurality of analytical models each associated with a corresponding partition of a gamut of the reproduction device, and wherein the recursively executing said steps comprises the online predicting for each of the plurality of analytical models. 15. A color printer comprising:an image reproduction system for generating an output image in response to an input signal;a sensor for detecting a sensor signal representative of the output image; and,a processor comprising:a parametric model of the image reproduction system for transforming the input signal into a model output;a real-time parameter adjustment algorithm;an inversion of the parametric model; and,a comparator for comparing the sensor signal with the model output, wherein based upon an output from the comparator, the real-time parameter adjustment algorithm computes an updated set of parameters for the parametric model and the inversion of the parametric model, and the model and inversion are updated in accordance with the updated parameters, for recursively transforming a subsequent input signal with the updated inversion and the updated model. 16. The printer as defined in claim 15 wherein the parametric model comprises a linear affine model. 17. The printer as defined in claim 15 wherein the adjustment algorithm comprises a recursive convergent least square estimation algorithm. 18. The printer as defined in claim 15 wherein the parametric model comprises a plurality of models partitioned to cumulatively represent a gamut of the printer. 19. The printer as defined in claim 15 wherein the parametric model comprises a partial quadratic model. 20. The printer as defined in claim 15 wherein the parametric model comprises a quadratic model. 21. The printer as defined in claim 15 wherein the parametric model comprises a cubic model. 22. A method of calibrating a color printer comprising updating a set of parameters for a dynamic parametric model representing color space transformations by the printer of an input signal to an output signal, wherein the printer includes a sensor for detecting a signal representative of printer color output, the method comprising steps of:producing an image with the printer in response to the input signal;sensing with the sensor the color signal representative of the printer color output;generating an output from the parametric model responsive to the input signal and comparable to the sensed color signal;determining a difference between the sensed signal and the parametric model output;adjusting the set of parameters in accordance with the difference for updating the parametric model;generating an inverse of the updated parametric model; and,transforming a subsequent input signal with the inverse of the updated parametric model whereby subsequent images produced by the printer are made in accordance with the inverse of the updated parametric model. 23. The method as defined in claim 22 wherein the adjusting comprises processing the difference through an estimation/prediction algorithm for computing an updated set of parameters for the dynamic parametric model. 24. The method as defined in claim 22 wherein the producing comprises applying a selected test target signal as the input signal.