In order to extract a foreground area more appropriately, an image processing apparatus for extracting a foreground area from an inputted image includes an image input unit that sets a first area and a second area different from the first area in the inputted image, a first extraction unit that extracts a foreground area from the first area, and a second extraction unit that extracts a foreground area from the second area by using an extraction method different from an extraction method used by the first extraction unit.

Methods and systems are described for registering a sports field to a video. Video of a live event may feature participants at a venue. A template of the venue, including virtual markings that represent real markings on the venue, may be obtained. A homographic transformation between an image plane and a ground plane may be determined by matching virtual markings to corresponding real markings captured in at least one frame of the video. The determined homographic transformation may be used in the automated analysis of sports statistics and in improving inserted annotations and visualizations.

A system for registering one or more cameras and/or creating an accurate three-dimensional (3D) model of a world space environment including back projecting at least one image from at least one of a plurality of camera views to the 3D model based on a set of existing camera parameters. The back projected image is added as a texture for the 3D model. This texture is automatically compared to one or more images from other camera views using a color space comparison of images to determine a set of differences or errors. The camera parameters and the 3D model are automatically adjusted to minimized the differences or errors. Over time, the parameters and the 3D model converge on a state that can be used to track moving objects, insert virtual graphics and/or perform other functions.

Sports and fitness applications for an autonomous unmanned aerial vehicle (UAV) are described. In an example embodiment, a UAV can be configured to track a human subject using perception inputs from one or more onboard sensors. The perception inputs can be utilized to generate values for various performance metrics associated with the activity of the human subject. In some embodiments, the perception inputs can be utilized to autonomously maneuver the UAV to lead the human subject to satisfy a performance goal. The UAV can also be configured to autonomously capture images of a sporting event and/or make rule determinations while officiating a sporting event.

Methods and systems are described for registering a sports field to a video. Video of a live event may feature participants at a venue. A template of the venue, including virtual markings that represent real markings on the venue, may be obtained. A homographic transformation between an image plane and a ground plane may be determined by matching virtual markings to corresponding real markings captured in at least one frame of the video. The determined homographic transformation may be used in the automated analysis of sports statistics and in improving inserted annotations and visualizations.

An example system for lightweight view dependent rendering is described herein. An example mobile device includes a display, instructions, and processor circuitry to execute instructions to extract a moving region within a first set of frames of volumetric video content captured by a plurality of real-world cameras facing a scene at different angles. The processor circuitry to generate a first billboard based on a cropped area of the frames in the first set of frames, the cropped area corresponding to the first moving region. The processor circuitry to, in response to user selection of a first view of the scene, cause presentation of the first billboard on the display. The processor circuitry to, in response to a change from the first view to a second view of the scene, cause a second billboard to replace the first billboard on the display.

A system for monitoring objects at sporting events or other types of events uses a wearable drone that has at least one camera or other sensor for capturing or otherwise sensing data. When the drone is to be used for monitoring, such as monitoring an object at a sporting event, the wearable drone may be detached from its user, and it may hover or otherwise fly within a certain position of an object to be monitored. While flying, the drone's sensor may be used to capture information, such as performance data or images, of the object during the sporting event.

A system and method for determining an absolute position of an object in an area is presented. The system includes a server having a processor, and a plurality of camera nodes coupled to the server. Each node includes a camera that acquires images of the object and area. The server receives image data from a camera, detects the object within an approximate location by image analysis techniques, and determines a relative position of the object in pixel coordinates. The processor then detects stationary markers proximate to the relative location of the object, determines an absolute position of the detected markers relative to known markers to define an absolute position of the marker, and determines an absolute location of the object in relation to the absolute location of the detected marker. This absolute position of the object is provided to an official to accurately locate the object in the area.

A system for registering one or more cameras and/or creating an accurate three-dimensional (3D) model of a world space environment including back projecting at least one image from at least one of a plurality of camera views to the 3D model based on a set of existing camera parameters. The back projected image is added as a texture for the 3D model. This texture is automatically compared to one or more images from other camera views using a color space comparison of images to determine a set of differences or errors. The camera parameters and the 3D model are automatically adjusted to minimized the differences or errors. Over time, the parameters and the 3D model converge on a state that can be used to track moving objects, insert virtual graphics and/or perform other functions.

An example system for lightweight view dependent rendering is described herein. The system includes at least one memory, and at least one processor to execute instructions to track a moving object within a first view of a scene, the first view captured by a first camera of a plurality of cameras. The plurality of cameras to capture video data of the scene from a plurality of angles. The at least one processor to extract a portion of the video data to obtain a cropped video corresponding to the moving object within the first view. The at least one processor to generate a billboard representation based on the cropped video. The at least one processor to cause the billboard representation to be presented on a display in front of a two-dimensional background.