Abstract: In various examples, detected object data representative of locations of detected objects in a field of view may be determined. One or more clusters of the detected objects may be generated based at least in part on the locations and features of the cluster may be determined for use as inputs to a machine learning model(s). A confidence score, computed by the machine learning model(s) based at least in part on the inputs, may be received, where the confidence score may be representative of a probability that the cluster corresponds to an object depicted at least partially in the field of view. Further examples provide approaches for determining ground truth data for training object detectors, such as for determining coverage values for ground truth objects using associated shapes, and for determining soft coverage values for ground truth objects.

Abstract: In various examples, detected object data representative of locations of detected objects in a field of view may be determined. One or more clusters of the detected objects may be generated based at least in part on the locations and features of the cluster may be determined for use as inputs to a machine learning model(s). A confidence score, computed by the machine learning model(s) based at least in part on the inputs, may be received, where the confidence score may be representative of a probability that the cluster corresponds to an object depicted at least partially in the field of view. Further examples provide approaches for determining ground truth data for training object detectors, such as for determining coverage values for ground truth objects using associated shapes, and for determining soft coverage values for ground truth objects.

Abstract: In various examples, detected object data representative of locations of detected objects in a field of view may be determined. One or more clusters of the detected objects may be generated based at least in part on the locations and features of the cluster may be determined for use as inputs to a machine learning model(s). A confidence score, computed by the machine learning model(s) based at least in part on the inputs, may be received, where the confidence score may be representative of a probability that the cluster corresponds to an object depicted at least partially in the field of view. Further examples provide approaches for determining ground truth data for training object detectors, such as for determining coverage values for ground truth objects using associated shapes, and for determining soft coverage values for ground truth objects.

Abstract: A panoramic image is generated from a plurality of source images. A panoramic analysis engine samples a first source image and a second source image included in the plurality of source images to generate a first proxy image and a second proxy image, respectively. The panoramic analysis engine samples inertial measurement information associated the two proxy images. The panoramic analysis engine detects a feature that is present in both the first proxy image and the second proxy image. The panoramic analysis engine blends the second proxy image into the first proxy image based on the inertial measurement information and a first position of the feature within the second proxy image relative to a second position of the feature within the first proxy image to generate a preview image. Finally, the panoramic analysis engine renders the preview image according to a first panoramic mode to generate a first partial display image.

Abstract: Embodiments of the present invention provide a novel solution that enables mobile devices to continuously track interesting subjects by creating dynamic visual models that can be used to detect and track subjects in-real time through total occlusion or even if a subject temporarily leaves the mobile device's field of view. Additionally, embodiments of the present invention use an online learning scheme that dynamically adjusts tracking procedures responsive to any appearance and/or environmental changes associated with an interesting subject that may occur over a period of time. In this manner, embodiments of the present invention can determine a more optimal focus position that allows movement by either the mobile device or the subject during the performance of auto-focusing procedures and also enables other camera parameters to properly calibrate (meter) themselves based on the focus position determined.

Abstract: A face beautification system and a method of face beautification. On embodiment of the face beautification system includes: (1) a coarse feature detector configured to generate an approximation of facial features in an image, (2) an edge-preserving filter configured to reduce distortions in the approximation, and (3) a feature enhancer operable to selectively filter a facial feature from said approximation and carry out an enhancement.

Abstract: An obstacle detection and tracking system identifies objects in the path of a vehicle equipped with the system and issues a visual, audible, and/or control system warning. The system includes a depth imaging system that acquires depth data from objects in the field of view of a detection zone encompassing at least a portion of the road to be driven upon. It is assumed most of the acquired data represents road plane information. Statistical analysis of the depth image data identifies in (X,Y,Z) space at least one plane of the road being driven, after which identification threshold normal heights above and below the road plane are defined. Imaged objects within the detection zone that are higher or lower than a threshold normal are deemed of potential concern and will generate a warning to the vehicle operator or vehicle.