Intelligent alignment system and method for color sensing devices
AssigneeVariable, Inc.
InventorsGeorge Yang Yu, Andrew Richard Temple, Jeff McGehee
Life of this patent
Estimating the term…
Abstract
A computer-implemented intelligent alignment method for a color sensing device (12) which initially obtains a raw sense value measured by the color sensing device and converts the raw sense value to a tristimulus value (S401). The color sensing device may be calibrated using a known set of colors (S402). Raw readings received from the color sensing device may be mapped to known tristimulus values (S403), and mapped values may be converted to a standard reference illuminant color space (S404). A client-server system may be provided in which a color sensing device and one or more servers collectively execute operations according to the present disclosure.
Claims16
A computer-implemented intelligent alignment method for a color sensing device, the method comprising:
- calibrating the color sensing device by
- obtaining raw sense values measured by the color sensing device with respect to each color in a known set of colors,
- converting the raw sense values to tristimulus values,
- providing a supervised training set associated with the calibrated color sensing device by
- mapping raw readings received from the color sensing device to known tristimulus values,
- converting the mapped values to a standard reference illuminant color space, the calibrated device readings in the standard color space defining unadjusted color readings,
- storing in a database a plurality of unadjusted color readings from the color sensing device and a plurality of reference scans from a specified reference device, and
- establishing a relationship between the color sensing device and the reference device wherein each of the plurality of unadjusted color readings has a corresponding reference scan with less than a maximum allowable ΔE in a same designated color set;
- receiving an input comprising a subsequent unadjusted color reading from the color sensing device;
- and
- generating to a user interface of a user computing device an output comprising an adjusted color reading by
- executing a K Nearest Neighbor (KNN) algorithm to search for K nearest unadjusted color readings in the database with respect to the received input unadjusted color reading, and
- returning a weighted average of the K corresponding reference scans.
A client-server system comprising:
- a color sensing device;
- and
- one or more servers configured to calibrate the color sensing device by
- obtaining raw sense values measured by the color sensing device with respect to each color in a known set of colors, and
- converting the raw sense values to tristimulus values;
- said one or more servers further configured to provide a supervised training set associated with the calibrated color sensing device by
- mapping raw readings received from the color sensing device to known tristimulus values,
- converting the mapped values to a standard reference illuminant color space, the calibrated device readings in the standard color space defining unadjusted color readings,
- storing in a database a plurality of unadjusted color readings from the color sensing device and a plurality of reference scans from a specified reference device,
- establishing a relationship between the color sensing device and the reference device wherein each of the plurality of unadjusted color readings has a corresponding reference scan with less than a maximum permissible ΔE in a same designated color set;
- wherein the one or more servers are further configured, responsive to receiving an input comprising a subsequent unadjusted color reading from the color sensing device, to generate an output comprising an adjusted color reading on a user interface associated with a display unit of a user computing device by
- executing a K Nearest Neighbor (KNN) algorithm to search for K nearest unadjusted color readings in the database with respect to the received input unadjusted color reading, and
- returning a weighted average of the K corresponding reference scans.
Dependencies are inferred from the claim text.