In various examples, a background of an object may be modified to generate a training image. A segmentation mask may be generated and used to generate an object image that includes image data representing the object. The object image may be integrated into a different background and used for data augmentation in training a neural network. Data augmentation may also be performed using hue adjustment (e.g., of the object image) and/or rendering three-dimensional capture data that corresponds to the object from selected views. Inference scores may be analyzed to select a background for an image to be included in a training dataset. Backgrounds may be selected and training images may be added to a training dataset iteratively during training (e.g., between epochs). Additionally, early or late fusion nay be employed that uses object mask data to improve inferencing performed by a neural network trained using object mask data.

Embodiments of the disclosure provide an angiography image determination method and an angiography image determination device. The method includes: obtaining a plurality of first images of a body part injected with a contrast medium; obtaining a plurality of corresponding second images by performing a first image preprocessing operation on each first image; obtaining a pixel statistical characteristic of each second image; finding a candidate image based on the pixel statistical characteristic of each second image; and finding a reference image corresponding to the candidate image among the plurality of first images.

Embodiments of the disclosure provide an angiography image determination method and an angiography image determination device. The method includes: obtaining a plurality of first images of a body part injected with a contrast medium; obtaining a plurality of corresponding second images by performing a first image preprocessing operation on each first image; obtaining a pixel statistical characteristic of each second image; finding a candidate image based on the pixel statistical characteristic of each second image; and finding a reference image corresponding to the candidate image among the plurality of first images.

Methods, systems, and apparatus, including computer programs encoded on a storage device, for electric grid asset detection are enclosed. An electric grid asset detection method includes: obtaining overhead imagery of a geographic region that includes electric grid wires; identifying the electric grid wires within the overhead imagery; and generating a polyline graph of the identified electric grid wires. The method includes replacing curves in polylines within the polyline graph with a series of fixed lines and endpoints; identifying, based on characteristics of the fixed lines and endpoints, a location of a utility pole that supports the electric grid wires; detecting an electric grid asset from street level imagery at the location of the utility pole; and generating a representation of the electric grid asset for use in a model of the electric grid.

Methods, systems, and apparatus, including computer programs encoded on a storage device, for electric grid asset detection are enclosed. An electric grid asset detection method includes: obtaining overhead imagery of a geographic region that includes electric grid wires; identifying the electric grid wires within the overhead imagery; and generating a polyline graph of the identified electric grid wires. The method includes replacing curves in polylines within the polyline graph with a series of fixed lines and endpoints; identifying, based on characteristics of the fixed lines and endpoints, a location of a utility pole that supports the electric grid wires; detecting an electric grid asset from street level imagery at the location of the utility pole; and generating a representation of the electric grid asset for use in a model of the electric grid.

Defects of interest and nuisance can be separated into different segments which enables detection of the defects of interest in only one segment. A region of an image can be segmented into a plurality of segments. A range attribute of the segments can be determined. Thresholding can be used to select one of the segments from the range attribute. The segment that is selected can be dilated.

To extend the working range of depth from defocus (DFD) particularly on small depth of field (DoF) images, DFD is performed on an image pair at multiple spatial resolutions and the depth estimates are then combined. Specific implementations construct a Gaussian pyramid for each image of an image pair, perform DFD on the corresponding pair of images at each level of the two image pyramids, convert DFD depth scores to physical depth values using calibration curves generated for each level, and combine the depth values from all levels in a coarse-to-fine manner to obtain a final depth map that covers the entire depth range of the scene.

To extend the working range of depth from defocus (DFD) particularly on small depth of field (DoF) images, DFD is performed on an image pair at multiple spatial resolutions and the depth estimates are then combined. Specific implementations construct a Gaussian pyramid for each image of an image pair, perform DFD on the corresponding pair of images at each level of the two image pyramids, convert DFD depth scores to physical depth values using calibration curves generated for each level, and combine the depth values from all levels in a coarse-to-fine manner to obtain a final depth map that covers the entire depth range of the scene.

A method of extracting and reconstructing court lines includes the steps of binarizing a court image of a court including court lines to form a binary image; performing horizontal projection for the binary image; searching for plural corners in the binary image and defining a court line range by the corners; forming plural linear segments from images within the court line range by linear transformation; defining at least one first cluster and at least one second cluster according to the characteristics of the linear segments and categorizing the linear segments into plural groups; taking an average of each group as a standard court line and creating a linear equation of the standard court line to locate the point of intersection of the standard court lines; and reconstructing the court lines according to the point of intersection. This method is capable of extracting the image of a portion of the court line from a dynamic or static image having a court line quickly to eliminate interference caused by noises coming from a portion other than the court line such as the background color, ambient brightness, people or advertisement, and reconstructing the court lines quickly and accurately to facilitate the determination of the boundary of a court line or the computation of data.

A method of extracting and reconstructing court lines includes the steps of binarizing a court image of a court including court lines to form a binary image; performing horizontal projection for the binary image; searching for plural corners in the binary image and defining a court line range by the corners; forming plural linear segments from images within the court line range by linear transformation; defining at least one first cluster and at least one second cluster according to the characteristics of the linear segments and categorizing the linear segments into plural groups; taking an average of each group as a standard court line and creating a linear equation of the standard court line to locate the point of intersection of the standard court lines; and reconstructing the court lines according to the point of intersection. This method is capable of extracting the image of a portion of the court line from a dynamic or static image having a court line quickly to eliminate interference caused by noises coming from a portion other than the court line such as the background color, ambient brightness, people or advertisement, and reconstructing the court lines quickly and accurately to facilitate the determination of the boundary of a court line or the computation of data.