Back to results
US 10,762,374 B2Granted
High accuracy image identification system
AssigneeBlinkfire Analytics, Inc.
InventorsStephen Joseph Olechowski, III, Nan Jiang, Alejandro Tatay de Pascual
Life of this patent
Estimating the term…
Abstract
An image identification system may identify key points on a known image and an unidentified image. One or more geometric shapes may be formed from the key points. A match between the unidentified image and the know image nay be determined by comparison of the respective geometric shapes.
Claims16
A method for image identification using a secure autonomous intelligent server, the method comprising:
- creating a database of known logos, the database comprising vertices of geometric shapes formed from the known logos;
- wherein the creating the database of known logos comprises:
- obtaining a known logo;
- creating blurred variations of the known logo, the blurred variations comprising the known logo portrayed in varying levels of blur;
- identifying key points on the known logo and the blurred variations;
- constructing a geometric shape from the key points of the known logo and the blurred variations;
- calculating vertices of the geometric shape of the known logo and the blurred variations;
- and
- storing the vertices of the geometric shape of the known logo and the blurred variations in the database and associating the vertices with the known logo and the blurred variations;
- obtaining an unidentified logo representing an object in a query image;
- identifying key points on the unidentified logo;
- constructing a geometric shape from the key points of the unidentified logo;
- calculating vertices of the geometric shape of the unidentified logo;
- matching the vertices of the geometric shape of the unidentified logo with the vertices of the geometric shape of at least one of the known logos and the blurred variations;
- constructing multiple geometric shapes comprising a plurality of triangles from the key points of the known logo and the blurred variations;
- calculating vertices of each of the multiple geometric shapes of the known logo and the blurred variations;
- constructing multiple geometric shapes comprising a plurality of triangles from the key points of the unidentified logo;
- calculating vertices of each of the multiple geometric shapes of the unidentified logo;
- and
- matching the vertices of two or more of the geometric shapes comprising the plurality of triangles of the unidentified logo with the vertices of the geometric shapes comprising the plurality of triangles of one of the known logos and the blurred variations.
17 more elements
A system for image identification, the system comprising:
- at least one processor;
- and
- a memory storing processor-executable instructions, wherein the at least one processor is configured to implement the following operations upon executing the processor-executable instructions:
- creating a database of known logos, the database comprising vertices of geometric shapes formed from the known logos;
- wherein the creating the database of known logos comprises:
- obtaining a known logo;
- creating blurred variations of the known logo, the blurred variations comprising the known logo portrayed in varying levels of blur;
- identifying key points on the known logo and the blurred variations;
- constructing a geometric shape from the key points of the known logo and the blurred variations;
- calculating vertices of the geometric shape of the known logo and the blurred variations;
- and
- storing the vertices of the geometric shape of the known logo and the blurred variations in the database and associating the vertices with the known logo and the blurred variations;
- obtaining an unidentified logo representing an object in a query image;
- identifying key points on the unidentified logo;
- constructing a geometric shape from the key points of the unidentified logo;
- calculating vertices of the geometric shape of the unidentified logo;
- matching the vertices of the geometric shape of the unidentified logo with the vertices of the geometric shape of at least one of the known logos and the blurred variations;
- constructing multiple geometric shapes comprising a plurality of triangles from the key points of the known logo and the blurred variations;
- calculating vertices of each of the multiple geometric shapes of the known logo and the blurred variations;
- constructing multiple geometric shapes comprising a plurality of triangles from the key points of the unidentified logo;
- calculating vertices of each of the multiple geometric shapes of the unidentified logo;
- and
- matching the vertices of two or more of the geometric shapes comprising the plurality of triangles of the unidentified logo with the vertices of the geometric shapes comprising the plurality of triangles of one of the known logos and the blurred variations.
20 more elements
A method for image identification using a secure autonomous intelligent server, the method comprising:
- creating a data base of known logos, the data base comprising vertices of geometric shapes formed from the known logos;
- wherein the creating the database of known logos comprises:
- obtaining a known logo;
- creating blurred variations of the known logo, the blurred variations comprising the known logo portrayed in varying levels of blur;
- wherein the creating the blurred variations of the known logo comprises:
- receiving an original image of a known logo;
- applying an image sample technique to the known logo;
- applying an image filtering technique to the known logo to smooth image pixels and generate multiple levels of blur;
- and
- receiving multiple synthetic templates of the known logo;
- identifying key points on the known logo and the blurred variations;
- constructing a geometric shape from the key points of the known logo and the blurred variations;
- calculating vertices of the geometric shape of the known logo and the blurred variations;
- and
- storing the vertices of the geometric shape of the known logo and the blurred variations in the database and associating the vertices with the known logo and the blurred variations;
- obtaining an unidentified logo representing an object in a query image;
- identifying key points on the unidentified logo;
- constructing a geometric shape from the key points of the unidentified logo;
- calculating vertices of the geometric shape of the unidentified logo;
- matching the vertices of the geometric shape of the unidentified logo with the vertices of the geometric shape of at least one of the known logos and the blurred variations;
- constructing multiple geometric shapes comprising a plurality of triangles from the key points of the known logo and the blurred variations;
- calculating vertices of each of the multiple geometric shapes of the known logo and the blurred variations;
- constructing multiple geometric shapes comprising a plurality of triangles from the key points of the unidentified logo;
- calculating vertices of each of the multiple geometric shapes of the unidentified logo;
- and
- matching the vertices of two or more of the geometric shapes comprising the plurality of triangles of the unidentified logo with the vertices of the geometric shapes comprising the plurality of triangles of one of the known logos and the blurred variations.
23 more elements
Dependencies are inferred from the claim text.