A moving object detection method and apparatus are disclosed. The method may include: obtaining a first image and a second image of a scene; determining a difference of the first image and the second image; performing a binarization operation on the difference of the first image and the second image, to generate a binary image; determining the number of pixels whose values are nonzero in each column of the binary image, to generate a column pixel histogram; and determining whether a moving object is present in the scene based on the column pixel histogram.

A digital medium environment is described to automatically generate a trimap and segment a digital image, independent of any user intervention. An image processing system receives an image and a low-resolution mask for the image, which provides a probability map indicating a likelihood that a pixel in the image mask corresponds to a foreground object in the image. The image processing system analyzes the image to identify content in the image's foreground and background portions, and adaptively generates a trimap for the image based on differences between the identified foreground and background content. By identifying content of the image prior to generating the trimap, the techniques described herein can be applied to a wide range of images, such as images where foreground content is visually similar to background content, and vice versa. Thus, the image processing system can automatically generate trimaps for images having diverse visual characteristics.

In example implementations, an image is identified. The image may be received. A radial histogram of each layer of the image is generated. The radial histogram of each layer of the image may be compared to at least one of a plurality of pre-generated radial histograms of a respective layer of a plurality of layers of a known image, wherein the plurality of layers is compressed into a consolidated set of bytes and the comparing is performed in parallel via the consolidate set of bytes. The image may be identified as a match of the known image when a match percentage of the radial histogram of the each layer of the image compared to the at least one of the plurality of pre-generated radial histograms of the respective layer of the plurality of layers of the known image is above a match threshold.

A data processing unit (101) processes input data using a neural network. A compression controlling unit (102) generates quantization information that defines quantization steps. An encoding unit (103) encodes network configuration information including parameter data which is quantized using the quantization steps determined by the compression controlling unit (102), and the quantization information, to generate compressed data.

A method for graphic image processing from existing source image files forming a database from which an image is extracted for said graphic processing operations. The method includes the step of implementing n (n1) saliency processing operations in order to form n saliency cards CSi, with i=1 to n; performing a linear combination $\frac{{}_{i=1}^N\mathrm{Ai}.\mathrm{SCi}}{\underset{i=1}{\overset{n}{.Math.}}\mathrm{Ai}}$
in order to obtain a single resulting saliency card; performing a first thresholding on the resulting saliency card; and vectorizing said thresholded card.

A system and method for determining a dimension of a target. The method includes: determining a camera parameter, the camera parameter including at least one of a focal length, a yaw angle, a roll angle, a pitch angle, or a height of one or more cameras; acquiring a first image and a second image of an target captured by the one or more cameras; generating a first corrected image and a second corrected image by correcting the first image and the second image; determining a parallax between a pixel in the first corrected image and a corresponding pixel in the second corrected image; determining an outline of the target; and determining a dimension of the target based at least in part on the camera parameter, the parallax, and the outline of the target.

An image processing apparatus includes a determination unit configured to determine a region of the image on which to perform character recognition processing, a decision unit configured to decide, based on a number of black pixels in contact with the region determined by the determination unit, whether to perform the character recognition processing on an expanded region obtained by expanding the region determined by the determination unit rather than on the region determined by the determination unit, and a character recognition unit configured to perform the character recognition processing on that region of the image decided by the decision unit.

Techniques are disclosed for creating a background model of a scene using both a pixel based approach and a context based approach. The combined approach provides an effective technique for segmenting scene foreground from background in frames of a video stream. Further, this approach can scale to process large numbers of camera feeds simultaneously, e.g., using parallel processing architectures, while still generating an accurate background model. Further, using both a pixel based approach and context based approach ensures that the video analytics system can effectively and efficiently respond to changes in a scene, without overly increasing computational complexity. In addition, techniques are disclosed for updating the background model, from frame-to-frame, by absorbing foreground pixels into the background model via an absorption window, and dynamically updating background/foreground thresholds.

The present disclosure provides a method and device for determining a handwriting similarity. The method includes: performing image processing on a handwriting image comprising a handwriting to be compared to obtain a first processed image and a second processed image having different handwriting features of the handwriting image; determining, based on the first processed image, a first feature vector indicating at least one first handwriting feature of the handwriting to be compared; determining, based on the second processed image, a second feature vector indicating at least one second handwriting feature of the handwriting to be compared; determining a handwriting feature vector of the handwriting to be compared based at least on the first feature vector and the second feature vector; and determining a similarity between the handwriting to be compared and a reference handwriting based on the handwriting feature vector and a reference handwriting feature vector of the reference handwriting.

A device receives, as part of a gesture translation service, data that depicts gestures, wherein the data is image data or multimedia data. The device converts a set of frames that include the data to another set of frames that include modified data identifying a grayscale or black and white depiction of the gestures and generates graphical representations of the gestures identified by the modified data. The device selects, using a matching technique, a graphical representation of corresponding gestures that matches or satisfies a threshold level of similarity with the graphical representations of the gestures identified by the modified data. The device identifies response data that is representative of the corresponding gestures based on the response data being stored in association with an identifier of the graphical representation that has been selected. The device provides the response data to be displayed, via an interface, as text data or audio data.