Semantic compression
First Claim
1. A method performed by a computing device to compress medical data, the method comprising:
- receiving, by the computing device, the medical data that represents one or more physical attributes;
identifying, by the computing device, a first region and a second region in the received medical data by;
characterizing a utility of different signal segments by analyzing one or more sets of diagnosis results for a first diagnosis, the utility corresponding to how important each of the signal segments is to the first diagnosis compared to other signal segments,identifying a unique combination of features in each of the signal segments, andderiving one or more algorithms for automatic feature identification using at least one of a pattern matching technique or a statistical technique to segment the received medical data into the first region and the second region such that each of the first region and the second region has a same utility value;
computing, by the computing device, a first entropy for the first region and a second entropy for the second region;
computing, by the computing device, a first utility value and a first feature for the first region, and a second utility value and a second feature for the second region, wherein the first utility value is representative of a relative measure of how important the first region is to the first diagnosis compared to the second region, and wherein the second utility value is representative of a relative measure of how important the second region is to the first diagnosis compared to the first region;
computing, by the computing device, a first rating for the first region and a second rating for the second region, wherein the first rating is a function of the first utility value and the first entropy, and the second rating is a function of the second utility value and the second entropy, and wherein the first rating is proportional to the first utility value and inversely proportional to the first entropy, and the second rating is proportional to the second utility value and inversely proportional to the second entropy;
determining, by the computing device, whether the first rating is greater than the second rating;
in response to determining that the first rating is greater than the second rating;
compressing, by the computing device, the first region of the received medical data using a first compression technique, andcompressing, by the computing device, the second region of the received medical data using a second compression technique,wherein the second compression technique is more lossy than the first compression technique; and
in response to determining that the first rating is less than or equal to the second rating;
compressing, by the computing device, the first region of the received medical data using the second compression technique, andcompressing, by the computing device, the second region of the received medical data using the first compression technique,wherein the first region and the second region are respectively compressed in accordance with a compression technique selected based on at least a measure of how important that region is to the first diagnosis and respective entropy for the first region and the second region.
5 Assignments
0 Petitions
Accused Products
Abstract
Technology for semantic compression is disclosed. In various embodiments, the technology receives data that represents one or more physical attributes sensed by one or more sensors; employs at least one pattern or statistical feature to identify a first region and a second region in the received data; computes a first utility and a first relevant feature for the first region, and a second utility and a second relevant feature for the second region; and identifies based on at least the first utility and the second utility a first compression method to apply to the first region and a second compression method to apply to the second region wherein the first and the second compression methods have different compression rates, different feature preservation characteristics, or both.
46 Citations
26 Claims
-
1. A method performed by a computing device to compress medical data, the method comprising:
-
receiving, by the computing device, the medical data that represents one or more physical attributes; identifying, by the computing device, a first region and a second region in the received medical data by; characterizing a utility of different signal segments by analyzing one or more sets of diagnosis results for a first diagnosis, the utility corresponding to how important each of the signal segments is to the first diagnosis compared to other signal segments, identifying a unique combination of features in each of the signal segments, and deriving one or more algorithms for automatic feature identification using at least one of a pattern matching technique or a statistical technique to segment the received medical data into the first region and the second region such that each of the first region and the second region has a same utility value; computing, by the computing device, a first entropy for the first region and a second entropy for the second region; computing, by the computing device, a first utility value and a first feature for the first region, and a second utility value and a second feature for the second region, wherein the first utility value is representative of a relative measure of how important the first region is to the first diagnosis compared to the second region, and wherein the second utility value is representative of a relative measure of how important the second region is to the first diagnosis compared to the first region; computing, by the computing device, a first rating for the first region and a second rating for the second region, wherein the first rating is a function of the first utility value and the first entropy, and the second rating is a function of the second utility value and the second entropy, and wherein the first rating is proportional to the first utility value and inversely proportional to the first entropy, and the second rating is proportional to the second utility value and inversely proportional to the second entropy; determining, by the computing device, whether the first rating is greater than the second rating; in response to determining that the first rating is greater than the second rating; compressing, by the computing device, the first region of the received medical data using a first compression technique, and compressing, by the computing device, the second region of the received medical data using a second compression technique, wherein the second compression technique is more lossy than the first compression technique; and in response to determining that the first rating is less than or equal to the second rating; compressing, by the computing device, the first region of the received medical data using the second compression technique, and compressing, by the computing device, the second region of the received medical data using the first compression technique, wherein the first region and the second region are respectively compressed in accordance with a compression technique selected based on at least a measure of how important that region is to the first diagnosis and respective entropy for the first region and the second region. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23)
-
-
24. A non-transitory computer-readable storage medium that stores computer-executable instructions that, in response to execution, cause a computing system to perform a method to compress medical data, the method comprising:
-
identifying a first region and a second region in the medical data, which represents one or more physical attributes, by; characterizing a utility of different signal segments by analyzing one or more sets of diagnosis results for a first diagnosis, the utility corresponding to how important each of the signal segments is to the first diagnosis compared to other signal segments, identifying a unique combination of features in each of the signal segments, and deriving one or more algorithms for automatic feature identification using at least one of a pattern matching technique or a statistical technique to segment the medical data into the first region and the second region such that each of the first region and the second region has a same utility value; computing a first entropy for the first region and a second entropy for the second region; computing a first utility value and a first feature for the first region, and a second utility value and a second feature for the second region, wherein the first utility value is representative of a relative measure of how important the first region is to the first diagnosis compared to the second region, and wherein the second utility value is representative of a relative measure of how important the second region is to the first diagnosis compared to the first region; computing a first rating for the first region and a second rating for the second region, wherein the first rating is a function of the first utility value and the first entropy, and the second rating is a function of the second utility value and the second entropy, the first rating being determined by an equation;
R1(X)=U1(X)/H1(X)wherein R1(X) is the first rating, U1(X) is the first utility value, and H1(X) is the first entropy, and the second rating being determined by an equation;
R2(X)=U2(X)/H2(X)wherein R2(X) is the second rating, U2(X) is the second utility value, and H2(X) is the second entropy; and determining whether the first rating is greater than the second rating; in response to determining that the first rating is greater than the second rating; compressing the first region of the medical data using a first compression technique, and compressing the second region of the medical data using a second compression technique, wherein the second compression technique is more lossy than the first compression technique; and in response to determining that the first rating is less than or equal to the second rating; compressing the first region of the medical data using the second compression technique, and compressing the second region of the medical data using the first compression technique, wherein the first region and the second region are respectively compressed in accordance with a compression technique selected based on at least a measure of how important that region is to the first diagnosis and respective entropy for the first region and the second region. - View Dependent Claims (26)
-
-
25. A system to compress medical data, the system comprising:
-
a component configured to receive the medical data that represents one or more physical attributes from a plurality of sensors; a component configured to identify a first region and a second region in the received medical data by; characterization of a utility of different signal segments by analysis of one or more sets of diagnosis results for a first diagnosis, the utility corresponding to how important each of the signal segments is to the first diagnosis compared to other signal segments, identification of a unique combination of features in each of the signal segments, and derivation of one or more algorithms for automatic feature identification using at least one of a pattern matching technique or a statistical technique to segment the received medical data into the first region and the second region such that each of the first region and the second region has a same utility value; a component configured to compute; a first entropy for the first region and a second entropy for the second region, a first utility value and a first feature for the first region, and a second utility value and a second feature for the second region, wherein the first utility value is representative of a relative measure of how important the first region is to the first diagnosis compared to the second region, and wherein the second utility value is representative of a relative measure of how important the second region is to the first diagnosis compared to the first region, and a first rating for the first region and a second rating for the second region, wherein the first rating is a function of the first utility value and the first entropy, and the second rating is a function of the second utility value and the second entropy, and wherein the first rating is proportional to the first utility value and inversely proportional to the first entropy, and the second rating is proportional to the second utility value and inversely proportional to the second entropy; a component configured to; compare different compression rates or different feature preservation characteristics for at least a first compression technique and a second compression technique after application of the first compression technique and the second compression technique to two or more sets of the received medical data; and determine whether there is another compression technique to utilize; a component configured to; in response to a determination that the second compression technique is more lossy than the first compression technique and there is no other compression technique to utilize; determine whether the first rating is greater than the second rating; and a component configured to; in response to a determination that the first rating is greater than the second rating; compress the first region of the received medical data using the first compression technique; and compress the second region of the received medical data using the second compression technique; and in response to a determination that the first rating is less than or equal to the second rating; compress the first region of the medical received data using the second compression technique; and compress the second region of the received medical data using the first compression technique, wherein the first region and the second region are respectively compressed in accordance with a compression technique selected based on at least a measure of how important that region is to the first diagnosis and respective entropy for the first region and the second region.
-
Specification