Systems and methods for compressing data

Patent metadata

Field	Value
Publication number	US-2018063538-A1
Application number	US-201615380519-A
Country	US
Kind code	A1
Filing date	Dec 15, 2016
Priority date	Aug 26, 2016
Publication date	Mar 1, 2018
Grant date	—

Abstract

A method of compressing data in the context of a decision-making task includes receiving raw data, analyzing the raw data to determine content of the raw data, and adjusting one or more one data compression parameters in a compression algorithm. The adjustment of the one or more compression parameters is based on the content of the raw data and a received decision-making task to produce a modified compression algorithm. The raw data is thereafter compressed using the modified compression algorithm and output as compressed data.

First claim

What is claimed is:

1 . A method of compressing data, comprising:
receiving raw data; analyzing the raw data to determine content; adjusting at least one compression parameter in a compression algorithm based on the content of the data and a received decision-making task to produce a modified compression algorithm; compressing the data using the modified compression algorithm; and outputting the data as compressed data.

2 . The method as recited in claim 1 , wherein the raw data includes data selected from a group including audio data, image data, and video data.

3 . The method as recited in claim 1 , further comprising analyzing the compressed data to issue a decision related to the decision-making task prior to receiving additional raw data.

4 . The method as recited in claim 1 , wherein the adjusted compression algorithm preserves with high fidelity features in the raw data that are relevant to the decision-making task.

5 . The method as recited in claim 1 , wherein the compression algorithm is an auto encoder and the compression parameter is initially established in a training stage.

6 . The method as recited in claim 5 , wherein the training stage adjusts at least one decision-making rule associated with the decision-making task based on the optimization of at least one of a reconstruction metric and a classification metric.

7 . The method as recited in claim 1 , further comprising receiving a decision-making task, wherein the decision-making task is received prior to receiving the raw data.

8 . The method as recited in claim 1 , further comprising receiving a decision-making task, wherein the decision-making task is received subsequent to receiving the raw data.

9 . A method of compressing image data in the context of a decision-making task, the method comprising:
receiving raw image data; analyzing the raw image data to determine content; adjusting at least one compression parameter in a compression algorithm based on the content of the raw image data and a received decision-making task to produce a modified compression algorithm; compressing the raw image data using the modified compression algorithm; and outputting the raw image data as compressed raw image data.

10 . A data compression system, comprising:
a data-receiving device; a controller operatively connected to the data-receiving device; and a non-transitory machine-readable memory disposed in communication with the controller and having instructions recorded thereon that, when read by the controller, cause the controller to:
receive raw data from the data-receiving device;
analyze the raw data to determine content of the raw data;
adjust at least one compression parameter in a compression algorithm based on the content of the data and a received decision-making task to produce a modified compression algorithm;
compress the data using the modified compression algorithm; and
output the data as compressed data.

11 . The system as recited in claim 10 , wherein the data-receiving device is a camera and the raw data includes image or video data.

12 . The system as recited in claim 10 , wherein the data-receiving device includes at least one of a global positioning system, a magnetometer, an accelerometer, a gyroscope, a network device, and a microphone, and wherein the raw data includes data descriptive of at least one of orientation, speed, position, acceleration, motion, audio, transaction, text, and speech.

13 . The system as recited in claim 10 , wherein the instructions cause the controller to adjust at least one decision-making rule relevant to the decision-making task based on the content of the compressed data, and wherein the compression algorithm and the decision-making rule are adjusted coincident with one another.

14 . The system as recited in claim 10 , wherein the modified compression algorithm preserves with high fidelity features in the raw data that are relevant to the decision-making task carried out by a decision-making algorithm.

15 . The system as recited in claim 14 , wherein the decision-making algorithm is a statistical classifier.

16 . The system as recited in claim 14 , wherein the statistical classifier includes at least one of a support vector machine, a decision tree, a k-nearest neighbor, a decision tree ensemble, and a neural network.

17 . The system as recited in claim 10 , wherein the compression algorithm includes an auto encoder, and wherein the parameter adjustment is done in a training stage of the autoencoder.

18 . The system as recited in claim 17 , wherein autoencoder includes one of a static autoencoder, a convolutional autoencoder, and a recurrent autoencoder.

19 . The system as recited in claim 14 , wherein the decision-making task is a classification task, and wherein the decision-making rule associated with the classification task is adjusted in a training stage based upon on optimization of at least one of a reconstruction metric and a classification metric.

20 . The system as recited in claim 11 , wherein the instructions cause the controller to receive a decision-making task prior to receiving the raw data.

Assignees

• Goodrich Corp

Inventors

• Bernal Edgar A
• Reddy Kishore K
• Giering Michael J

Classifications

• G06V20/13 Primary

  Satellite images · CPC title

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• H04N23/69

  Control of means for changing angle of the field of view, e.g. optical zoom objectives or electronic zooming · CPC title

• G06F18/24

  Classification techniques · CPC title

• G06V10/25

  Determination of region of interest [ROI] or a volume of interest [VOI] · CPC title