Detection of hand gestures using gesture language discrete values

Abstract

Computer implemented method for detecting a hand gesture of a user, comprising:(a) Receiving sequential logic models each representing a hand gesture. The sequential logic model maps pre-defined hand poses and motions each represented by a hand features record defined by discrete hand values each indicating a state of respective hand feature.(b) Receiving a runtime sequence of runtime hand datasets each defined by discrete hand values scores indicating current state hand features of a user's moving hand which are inferred by analyzing timed images depicting the moving hand.(c) Submitting the runtime hand datasets and the pre-defined hand features records in SSVM functions to generate estimation terms for the runtime hand datasets with respect to the hand features records.(d) Estimating which of the hand gestures best matches the runtime sequence depicted in the timed images by optimizing score functions using the estimation terms for the runtime hand datasets.

What is claimed is: 
     
         1 . A computer-implemented method comprising:
 obtaining a logic model representing a hand gesture, the logic model correlating to at least one of: a pre-defined hand pose or a pre-defined hand motion;   obtaining a runtime sequence for a moving hand, the runtime sequence comprising a runtime hand dataset that includes discrete motion values indicative of one or more of: a palm pose feature, a finger flexion feature, a finger tangency condition feature, a finger relative location condition feature;   generating an estimation term based on the discrete motion values; and   estimating that the hand gesture identifies the moving hand using the estimation term.   
     
     
         2 . The computer-implemented method of  claim 1 , further comprising detecting a multimodal act that comprises the hand gesture. 
     
     
         3 . The computer-implemented method of  claim 1 , wherein the runtime hand dataset is estimated as at least one of: a hand pose that is not pre-defined or a hand motion that is not pre-defined. 
     
     
         4 . The computer-implemented method of  claim 1 , further comprising implementing a structured support vector machine (SSVM) function to generate the estimation term at least by applying a parametric function to the runtime hand dataset, wherein the parametric function simulates a discrete hand value. 
     
     
         5 . The computer-implemented method of  claim 1 , wherein estimating that the hand gesture identifies the moving hand comprises analyzing a pairwise term. 
     
     
         6 . The computer-implemented method of  claim 1 , further comprising implementing a structured support vector machine (SSVM) function that receives the runtime hand dataset as an input. 
     
     
         7 . The computer-implemented method of  claim 1 , wherein the logic model is represented as a finite state machine (FSM) having a state correlating to a pre-defined hand features record, wherein the FSM is configured to be augmented with a score function. 
     
     
         8 . The computer-implemented method of  claim 1 , further comprising implementing a structured support vector machine (SSVM) function that selects a context registered hand gesture of the hand gesture. 
     
     
         9 . The computer-implemented method of  claim 1 , wherein the discrete motion values are represented by a Boolean formula that is defined in the form of a Conjunctive Normal Form (CNF). 
     
     
         10 . The computer-implemented method of  claim 1 , wherein estimating that the hand gesture identifies the moving hand comprises submitting the runtime hand dataset to a structured support vector machine (SSVM) function together with a pre-defined hand features record. 
     
     
         11 . A system comprising:
 a memory storing code;   a processor coupled to the memory for executing the code to:
 obtain a logic model representing a hand gesture, the logic model correlating to at least one of: a pre-defined hand pose or a pre-defined hand motion; 
 obtain a runtime sequence for a moving hand, the runtime sequence comprising a runtime hand dataset that includes discrete motion values indicative of one or more of: a palm pose feature, a finger flexion feature, a finger tangency condition feature, or a finger relative location condition feature; 
 generate an estimation term based on the discrete motion values; and 
 estimate that the hand gesture identifies the moving hand using the estimation term. 
   
     
     
         12 . The system of  claim 11 , wherein the logic model is represented as a finite state machine (FSM) having a state correlating to a pre-defined hand features record, said FSM configured to be augmented with a score function. 
     
     
         13 . The system of  claim 11 , wherein the code implements a structured support vector machine (SSVM) that selects a context registered hand gesture of the hand gesture. 
     
     
         14 . The system of  claim 11 , wherein the discrete motion values are represented by a Boolean formula that is defined in the form of a Conjunctive Normal Form (CNF). 
     
     
         15 . The system of  claim 11 , wherein estimating that the hand gesture identifies the moving hand comprises submitting the runtime hand dataset to a structured support vector machine (SSVM) function. 
     
     
         16 . The system of  claim 11 , wherein the logic model is represented as an FSM, a state of the FSM correlates to the pre-defined hand features record, and the FSM is augmented with a score function. 
     
     
         17 . The system of  claim 11 , wherein the code implements a structured support vector machine (SSVM) function that generates the estimation term at least by applying a parametric function to the runtime hand dataset, wherein the parametric function simulates a discrete hand value. 
     
     
         18 . The system of  claim 11 , wherein the code detects a multimodal act that comprises at least one of: the hand gesture or a non-gesture user interaction, wherein the non-gesture user interaction comprises one or more of: a text input, a visual element selection, a tactile input, a voice input. 
     
     
         19 . A computer-storage memory device embodied with executable instructions comprising:
 first program instructions to obtain a logic model representing a hand gesture, the logic model correlating to at least one of: a pre-defined hand pose or a pre-defined hand motion;   second program instructions to obtain a runtime sequence for a moving hand, the runtime sequence comprising a runtime hand dataset that includes discrete motion values indicative of one or more of: a palm pose feature, a finger flexion feature, a finger tangency condition feature, a finger relative location condition feature;   third program instructions to generate an estimation term based on the discrete motion values; and   fourth program instructions to estimate that the hand gesture identifies the moving hand using the estimation term.   wherein the first, second, third and fourth program instructions are executed by a computerized processor.   
     
     
         20 . The computer-storage memory device of  claim 19 , wherein the finger tangency condition feature defines a touch condition of two fingers.