Sign-language automated teller machine

A method for an automated teller machine comprising:

• receiving video data from a camera, the camera configured to capture the video data from a user, the video data comprising at least one hand movement generated by the user;
• splitting the video data into a first image, a second image, a third image, and a fourth image;
• splitting the first image into a first set of features;
• forming a first feature vector based on the first set of features;
• splitting the second image into a second set of features;
• forming a second feature vector based on the second set of features;
• splitting the third image into a third set of features;
• forming a third feature vector based on the third set of features;
• splitting the fourth image into a fourth set of features;
• forming a fourth feature vector based on the fourth set of features;
• determining a first distance between the first feature vector and the second feature vector;
• determining that the first distance is within a set tolerance;
• in response to determining that the first distance is within the set tolerance:
• comparing the first feature vector to feature vectors stored in a first collection of feature vectors, each feature vector of the first collection assigned to a first type of sign-language pattern comprising a motionless sign-language hand position;
• determining, based on the comparison of the first feature vector to the feature vectors stored in the first collection of feature vectors, that the first feature vector corresponds to a first motionless sign-language hand position assigned to a first word or a first phrase;
• and
• storing the first word or the first phrase in an input phrase;
• determining a second distance between the second feature vector and the third feature vector;
• determining a third distance between the third feature vector and the fourth feature vector;
• determining that the second distance is not within the set tolerance;
• determining that the third distance is not within the set tolerance;
• in response to determining that the second distance is not within the set tolerance and that the third distance is not within the set tolerance:
• comparing the third feature vector and the fourth feature vector to groupings of feature vectors stored in a second collection of groupings of feature vectors, each grouping of feature vectors stored in the second collection comprising at least two feature vectors and assigned to a second type of sign-language pattern comprising a moving sign-language hand gesture;
• determining, based on the comparison of the third feature vector and the fourth feature vector to the groupings of feature vectors in the second collection of groupings of feature vectors, that the third feature vector and the fourth feature vector comprise a first moving sign-language hand gesture assigned to a second word or a second phrase;
• and
• storing the second word or the second phrase in the input phrase;
• determining that the input phrase represents a user request;
• and
• in response to determining that the input phrase represents the user request, processing the user request.

A system comprising:

• an automated teller machine;
• a camera configured to capture video data from a user, the video data comprising at least one hand movement generated by the user;
• a storage element configured to store a dictionary of sign-language patterns, each sign-language pattern assigned to a word or a phrase, the dictionary comprising:
• a first collection of feature vectors, each feature vector of the first collection assigned to a first type of sign-language pattern comprising a motionless sign-language hand position;
• and
• a second collection of groupings of feature vectors, each grouping of feature vectors of the second collection comprising at least two feature vectors and assigned to a second type of sign-language pattern comprising a moving sign-language hand gesture;
• a processing element, communicatively coupled to the automated teller machine, the camera, and the memory, the processing element operable to:
• receive the video data from the camera;
• split the video data into a first image, a second image, a third image, and a fourth image;
• split the first image into a first set of features;
• form a first feature vector based on the first set of features;
• split the second image into a second set of features;
• form a second feature vector based on the second set of features;
• split the third image into a third set of features;
• form a third feature vector based on the third set of features;
• split the fourth image into a fourth set of features;
• form a fourth feature vector based on the fourth set of features;
• determine a first distance between the first feature vector and the second feature vector;
• determine that the first distance is within a set tolerance;
• in response to determining that the first distance is within the set tolerance:
• compare the first feature vector to the feature vectors in the first collection of feature vectors;
• determine, based on the comparison of the first feature vector to the feature vectors in the first collection of feature vectors, that the first feature vector corresponds to a first motionless sign-language hand position assigned to a first word or a first phrase;
• and
• store the first word or the first phrase in an input phrase;
• determine a second distance between the second feature vector and the third feature vector;
• determine a third distance between the third feature vector and the fourth feature vector;
• determine that the second distance is not within the set tolerance;
• determine that the third distance is not within the set tolerance;
• in response to determining that the second distance is not within the set tolerance and that the third distance is not within the set tolerance:
• compare the third feature vector and the fourth feature vector to the groupings of feature vectors in the second collection of groupings of feature vectors;
• determine, based on the comparison of the third feature vector and the fourth feature vector to the groupings of feature vectors in the second collection of groupings of feature vectors, that the third feature vector and the fourth feature vector comprise a first moving sign-language hand gesture assigned to a second word or a second phrase;
• and
• store the second word or the second phrase in the input phrase;
• determine that the input phrase represents a user request;
• and
• in response to determining that the input phrase represents a user request:
• send the input phrase to a device belonging to the user;
• receive a response from the user indicating that the input phrase is correct;
• and
• process the user request.