Method and apparatus for automatic search for relevant picture data sets

• US 6,735,577 B2
• Filed: 07/02/2001
• Issued: 05/11/2004
• Est. Priority Date: 07/03/2000
• Status: Expired due to Fees

• Current Assignee: Dr. Ing. Horst Zuse
• Original Assignee: Dr.-Ing. Horst Zuse
• Inventors: Zuse, Horst
• Primary Examiner(s): Davis, George B.
• Application Number:
  US09/898,755
• Time in Patent Office: 1,044 Days
• Field of Search: 706/12
• US Class Current: 706/12
• CPC Class Codes: G06F 16/532 Query formulation, e.g. graphical querying

Abstract

An apparatus is configured to automatically search for relevant picture data sets from a quantity of n (n≧2) picture data sets electronically stored in a memory device. Picture attributes for each of the n picture data sets is stored electronically in the memory device, and the n picture data sets as well as the stored picture attributes are adapted to be processed electronically by processor device. A training quantity including the electronically stored picture attributes of picture data sets which were selected by a user is utilized in the context of a machine learning process carried out automatically with the aid of the processor device to determine a decision function f. Subsequently, system relevancies of the picture data sets are calculated, based on the decision function f. Selected picture data sets then are provided in the context of the search result to be output in response to the respective system relevance.

Claims

1. A method of automatically searching for relevant picture data sets in a quantity of n (n≧

2) picture data sets electronically stored in a memory device, picture attributes for each of the n picture data sets being stored electronically in the memory device, and the n picture data sets as well as the stored picture attributes being adapted to be processed electronically by a processor, said method comprising;

(a) providing a first selection of picture data sets from the n picture data sets using the processor to be output by a display device;

(b) outputting several of the picture data sets of the first selection of picture data sets using the display device;

(c) electronically recording a respective evaluation by a user for at least one relevant picture data set of the plurality of picture data sets output according to (b); and

(d) providing a second selection of m (m≦

n) picture data sets from the n picture data sets in a sequence which depends on a respective system relevance of the m picture data sets to be output by the display device;

a machine learning process being carried out for electronically determining a decision function f to provide the second selection of the m picture data sets;

the picture attributes electronically stored in the memory device for the at least one relevant picture data set constituting a training quantity for the machine learning process;

the respective system relevance being determined for k (k≧

m) picture data sets using the decision function f and the respective electronically stored picture attributes; and

the k picture data sets comprising at least part of the m picture data sets of the second selection.

2. The method as claimed in claim 1, wherein the first selection of picture data sets from the n picture data sets is determined by device of a catchword search.

3. The method as claimed in claim 1, wherein the electronic recording of the respective evaluation by the user for the at least one relevant picture data set according to (c) comprises recording an actuation of an electronic selector device.

4. The method as claimed in claim 3, wherein the selector device is a mouse device configured to cooperate with the display device.

5. The method as claimed in claim 1, wherein the respective evaluation of the user for the at least one relevant picture data set according to (c) is recorded electronically as a binary evaluation so that each picture data set evaluated is adapted to be recorded as a relevant picture data set and each picture data set not evaluated is adapted to be recorded as a non-relevant picture data set.

6. The method as claimed in claim 1, wherein the decision function f is determined in the context of the machine learning process so that the electronic determination of the respective system relevance is optimized as regards the time period it requires.

7. The method as claimed in claim 1, wherein the machine learning process is carried out as a perceptron learning process.

8. The method as claimed in claim 1, wherein the picture attributes stored comprise catchwords suitable for electronic evaluation by the processor.

9. A picture search apparatus, comprising:

a memory device configured to electronically store n (n≧

2) picture data sets and respective picture attributes each associated respectively with the n picture data sets;

a display device configured to output a first selection of picture data sets from the n picture data sets;

a recording device configured to electronically record a respective evaluation by a user for at least one relevant picture data set of the first selection of picture data sets output;

a processor configured to automatically carry out a machine learning process to determine a decision function f in consideration of the picture attributes stored electronically in the memory device for the at least one relevant picture data set, and to electronically determine a respective system relevance for m (m≦

n) picture data sets of the n picture data sets stored electronically in the memory device with the aid of the decision function f and the respective electronically stored picture attributes; and

a device for providing a second selection of picture data sets in a sequence which depends on the respective system relevance to be output by the output by the display device, the second selection of picture data sets comprising k (k≧

m) picture data sets of the m picture data sets.

10. The picture search apparatus as claimed in claim 9, further comprising a search module configured to carry out a catchword search to search for the first selection of picture data sets.

11. A computer readable program product comprising computer program code for automatically searching for relevant picture data sets in a quantity of n (n≧

2) picture data sets electronically stored in a memory device, picture attributes for each of the n picture data sets being stored electronically in the memory device, and the n picture data sets as well as the stored picture attributes being adapted to be processed electronically by a processor, said computer program code providing for execution of a method comprising;

(a) providing a first selection of picture data sets from the n picture data sets using the processor to be output by a display device;

(b) outputting several of the picture data sets of the first selection of picture data sets using the display device;

(c) electronically recording a respective evaluation by a user for at least one relevant picture data set of the plurality of picture data sets output according to (b); and

(d) providing a second selection of m (m≦

n) picture data sets from the n picture data sets in a sequence which depends on a respective system relevance of the m picture data sets to be output by the display device;

a machine learning process being carried out for electronically determining a decision function f to provide the second selection of the m picture data sets;

the picture attributes electronically stored in the memory device for the at least one relevant picture data set constituting a training quantity for the machine learning process;

the respective system relevance being determined for k (k≧

m) picture data sets using the decision function f and the respective electronically stored picture attributes; and

the k picture data sets comprising at least part of the m picture data sets of the second selection.