Abstract: A fingerprinting solution that uses neural network (NN) based trained Machine Learning (ML) modules in combination with traditional image processing for contactless fingerprint capture, liveness detection to rule out fake fingers, and fingerprint matching using a portable handheld device with integrated camera, thereby eliminating the need for a special device dedicated for fingerprinting. The trained NN modules detect the size and direction of fingers in the captured image, check if fingers are reversed in the image (thereby making nails visible), check if the thumb of the correct hand is captured, and generate fixed-length fingerprint templates for subsequent matching of fingerprints. Three dimensional (3D) depth map of a finger is used to bring the fingerprint resolution to 500 dpi and eliminate distortion caused by the curvature of the finger shape to improve accuracy while scaling and flattening a fingerprint image.

Abstract: A passive authentication method includes, in response to receiving a requested action from a first user, obtaining a set of sensor data and categorizing first sensor data of the set of sensor data into a first modality of a set of modalities. The method includes, for the first modality of the set of modalities, determining a distance value by applying a first modality model to the first sensor data and comparing the distance value to a first verified value of the first user for the first modality. The method includes, based on the comparison, determining a first authentication decision of the distance value. The method includes, in response to the first authentication decision indicating the first sensor data corresponds to the first user, performing the requested action.

Abstract: A method is provided for classifying objects. The method detects objects in one or more images. The method tags each object with multiple features. Each feature describes a specific object attribute and has a range of values to assist with a determination of an overall quality of the one or more images. The method specifies a set of training examples by classifying the overall quality of at least some of the objects as being of an acceptable quality or an unacceptable quality, based on a user's domain knowledge about an application program that takes the objects as inputs. The method constructs a plurality of first-level classifiers using the set of training examples. The method constructs a second-level classifier from outputs of the first-level automatic classifiers. The second-level classifier is for providing a classification for at least some of the objects of either the acceptable quality or the unacceptable quality.

Abstract: Methods and systems for image clustering include matching a new image to a representative image of a cluster. The new image is set as a representative of the cluster with a first time limit. The new image is set as a representative of the cluster with a second time limit, responsive to a determination that the new image has matched at least one incoming image during the first time limit.

Abstract: A passive authentication method includes, in response to receiving a requested action from a first user, obtaining a set of sensor data and categorizing first sensor data of the set of sensor data into a first modality of a set of modalities. The method includes, for the first modality of the set of modalities, determining a distance value by applying a first modality model to the first sensor data and comparing the distance value to a first verified value of the first user for the first modality. The method includes, based on the comparison, determining a first authentication decision of the distance value. The method includes, in response to the first authentication decision indicating the first sensor data corresponds to the first user, performing the requested action.

Abstract: Methods and systems for image clustering include matching a new image to a representative image of a cluster. The new image is set as a representative of the cluster with a first time limit. The new image is set as a representative of the cluster with a second time limit, responsive to a determination that the new image has matched at least one incoming image during the first time limit.

Abstract: A method is provided for classifying objects. The method detects objects in one or more images. The method tags each object with multiple features. Each feature describes a specific object attribute and has a range of values to assist with a determination of an overall quality of the one or more images. The method specifies a set of training examples by classifying the overall quality of at least some of the objects as being of an acceptable quality or an unacceptable quality, based on a user's domain knowledge about an application program that takes the objects as inputs. The method constructs a plurality of first-level classifiers using the set of training examples. The method constructs a second-level classifier from outputs of the first-level automatic classifiers. The second-level classifier is for providing a classification for at least some of the objects of either the acceptable quality or the unacceptable quality.