Predicate Logic based Image Grammars for Complex Visual Pattern Recognition
1 Assignment
0 Petitions
Accused Products
Abstract
First order predicate logics are provided, extended with a bilattice based uncertainty handling formalism, as a means of formally encoding pattern grmmars, to parse a set of image features, and detect the presence of different patterns of interest implemented on a processor. Information from different sources and uncertainties from detections, are integrated within the bilattice framework. Automated logical rule weight learning in the computer vision domain applies a rule weight optimization method which casts the instantiated inference tree as a knowledge-based neural network, to converge upon a set of rule weights that give optimal performance within the bilattice framework. Applications are in (a) detecting the presence of humans under partial occlusions and (b) detecting large complex man made structures in satellite imagery (c) detection of spatio-temporal human and vehicular activities in video and (c) parsing of Graphical User Interfaces.
79 Citations
28 Claims
-
1-8. -8. (canceled)
-
9. A method for a detection of a visual pattern from image data of a scene, comprising:
-
obtaining image data of the scene by a processor; processing image data with one or more detectors implemented on the processor to detect one or more features; defining a pattern grammar to describe the visual pattern by encoding knowledge about contextual cues, scene geometry and visual pattern constraints; encoding the pattern grammar using a plurality of first order predicate logic based rules; executing by the processor the plurality of first order predicate logic based rules to detect the visual pattern based on a presence or an absence of the one or more features; and generating by the processor data related to; a justification that the plurality of first order predicate logic based rules detected the visual pattern; a location and time stamp in the scene where the visual pattern occurs, and a measure of uncertainty related to the detection of the visual pattern. - View Dependent Claims (10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25)
-
-
26. A system for a detection of a visual pattern from image data of scene, comprising:
-
a memory to store and retrieve data and instructions; a processor that accesses the memory to execute instructions to; obtain image data of the scene; process image data with one or more detectors to detect one or more features; define a pattern grammar to describe the visual pattern in a language that applies the detected one or more features; encode a pattern grammar that describes the visual pattern related to the detected one or more features in a plurality of first order predicate logic based rules; execute the plurality of first order predicate logic based rules to detect the visual pattern based on a presence or an absence of the one or more features; and generate data related to; a justification that the plurality of first order predicate logic based rules detected the visual pattern; a location and time stamp in the scene where the visual pattern occurs, and a measure of uncertainty related to the detection of the visual pattern. - View Dependent Claims (27, 28)
-
Specification