Abstract: Disclosed herein are systems and methods for using machine learning for image-based spoof detection. One embodiment takes the form of a method that includes obtaining an input-data set that includes a plurality of images captured of a biometric-authentication subject by a plurality of cameras of a camera system. The method also includes inputting the input-data set into a trained machine-learning module, and processing the input-data set using the machine-learning module to obtain, from the machine-learning module, a spoof-detection result for the biometric-authentication subject. The method also includes outputting the spoof-detection result for the biometric-authentication subject.

Abstract: Of a plurality of learning models learned to output geometric information corresponding to a captured image, a learning model corresponding to a setting region is acquired.

Abstract: In a vehicle surroundings recognition apparatus that recognizes a specific target around a vehicle from an image captured by an imaging unit, a shadow detection unit is configured to detect a shadow region based on a difference, between a plurality of elements constituting the image, in intensity of a specific color component included in colors represented by the plurality of elements and a difference in luminance between the plurality of elements, the shadow region being a region in the image in which a shadow is cast on a surface of the road. A feature point detection unit is configured to detect feature points in the image. A recognition unit is configured to recognize the target based on the shadow region detected by the shadow detection unit and a group of feature points detected by the feature point detection unit.

Abstract: A white balance method for an image and a terminal device are provided. The image includes a face region and a background region. The method includes extracting a color characteristic from the background region in the image, wherein the color characteristic is configured for indicating color distribution of the background region; matching the color characteristic with at least one application condition of a white balance algorithm for the face region, wherein the white balance algorithm is configured to adjust the face region in the image to be in a skin color; and performing a white-balance processing on the face region in the image according to the white balance algorithm in response to the color characteristic being matched with a corresponding one of the at least one application condition.

Abstract: In example embodiments, a computing system is capable of (a) receiving image data that represents a scene that was scanned by a detection device of a security screening system, (b) based at least on the image data, at least one neural network, and a confidence threshold defined for the security screening system, determining that the image data includes an identified item that has been deemed to be of interest, (c) based at least on determining that the image data includes the identified item that has been deemed to be of interest and one or more security parameters for the security screening system, determining that the identified item is deemed to be a security interest for the security screening system, and (d) based at least on determining that the identified item is deemed to be a security interest for the security screening system, presenting a visualization corresponding to the identified item.

Abstract: The present disclosure provides a line segment detection method, apparatus, device, and a computer-readable storage medium. Said method includes: extracting image features of an image to be processed, the image features including an image gradient angle and an image gradient amplitude; determining a plurality of seed points based on the extracted image features; for each seed point of the plurality of seed points, determining a current connected region of respective seed point according to a region growth rule, wherein the region growth rule needs to satisfy both a gradient amplitude growth rule and a gradient angle growth rule simultaneously; and performing line segment fitting on line segments within the current connected region of respective seed point.

Abstract: A histogram-based method for auto segmentation of integrated circuit structures is disclosed. The method includes an auto-segmentation process/algorithm, which works on the histogram of the SEM image and does not try to model the noise sources or the features. The auto-segmentation process/algorithm extracts the number of peaks in the histogram from low magnification SEM images or SEM images not necessarily having high quality images, significantly simplifies the traditionally lengthy and expensive IC reverse engineering efforts. Hence, the size of the image does not affect the final segmentation. The auto-segmentation process/algorithm performs the steps of: extract a first histogram from the first SEM image; identifying boundaries of the plurality of structural elements in the IC based at least in part on an output of the first histogram; and auto-segmenting the first SEM image into the plurality of structural elements.

Abstract: Provided is an a priori constraint and outlier suppression based image deblurring method. A convolution model is used for fitting a blurring process of a clear image and then the blurred image I is restored, so that the purpose of image deblurring is achieved. The method comprises an evaluation process of the significant structure of a blurred image, a process of blurring kernel estimation and outlier suppression, and a process of restoring the blurred image by non-blind deconvolution. A structure in the blurred image is obtained by use of L0 norm constraint and heavy-tailed a priori information. The L0 norm constraint is used to evaluate the blurring kernel. The evaluated blurring kernel is subjected to outlier suppression. The final restored image is obtained by using a non-blind deconvolution algorithm. The present invention can prominently improve the restoration level of the blurred image.

Abstract: A method for image dehazing based on adaptively improved linear global atmospheric light of a dark channel. A haze image in haze weather is first obtained, a variation angle of atmospheric light of the image is obtain through calculating a slope of a connection line between a center and a center of gravity of a binary image of the image, a linear atmospheric light map that varies regularly along a variation direction of the atmospheric light is obtains, a dehazed image is solved through an atmospheric scattering model, and then a processed haze image taken in the haze weather is output.

Abstract: An image measuring method performed with an image measuring device measuring a dimension of a measured object from an image of the measured object captured by an image capturer. The method executes a standard reference object measurement measuring a dimension of the standard reference object with the image measuring device; a standard reference dimension input inputting a dimension of the standard reference object specified by a device other than the image measuring device; a preset value calculation calculating a preset value from the dimension of the measured standard reference object and from a dimension of the standard reference object measured by a predetermined measurement tool; a measurement measuring a dimension of a measured object other than the standard reference object using the image measuring device; and a correction correcting the dimension of the measured object other than the standard reference object measured by the image measuring device.

Abstract: Methods and systems for training a neural network to distinguish between text documents and image documents are described. A corpus of text and image documents is obtained. A page of a text document is scanned by shifting a text window to a plurality of locations. In accordance with a determination that the text in the window at a respective location meets text line criteria, the text in the window is stored as a respective text snippet. A plurality of image windows are superimposed over at least one page of an image document. In accordance with a determination that the content of a respective image window meets image criteria, content of the image window is stored as a respective image snippet. The respective text snippet and the respective image snippet are provided to a classifier.

Abstract: The present disclosure relates to an image processing device, an imaging device, and an image processing method to superimpose a complementary image in such a manner that an appearance becomes more natural. From an image acquired by imaging of an arbitrary subject, a removal region selecting unit selects a region, in which an intended object to be removed is included, as a removal region with reference to distance information that is acquired when the image is imaged and that indicates a distance to the subject. Based on the distance information, a complementary region specifying unit specifies a complementary region to complement the removal region from a region including a background of the object to be removed in the image. An image superimposition processing unit superimposes a complementary image generated from the complementary region on a place from which the removal region is deleted in the image.

Abstract: A method for automated gear contact pattern verification includes applying a colored powder to at least one gear to be meshed to form contacts on at least one gear tooth of the at least one gear, and using the smartphone, capturing images of the at least one gear tooth. A matching algorithm run on the smartphone may include identifying bounding points of a yellow portion of the at least one gear tooth, the bounding points including a midpoint of the yellow portion; identifying a gear mesh area between the bounding points and a midpoint of the gear mesh area; and determining a deviation in the midpoints of the yellow portion and gear mesh area, wherein the deviation may between approximately 25% and 80%. The test results are displayed on the smartphone, the test results including at least an indication of pass or fail.

Abstract: An image processing apparatus configured to divide an image into frequency bands and reduce noise, in which a first image includes a high-band frequency component, and a second image includes a low-band frequency component, includes: a first detecting unit configured to detect an edge in at least one of the first image and the second image; a second detecting unit configured to detect a low-contrast edge that is of a lower contrast than a contrast of the edge detected by the first detecting unit in the at least one of the first image and the second image; a compositing unit configured to composite the first image and the second image using a weighting corresponding to the edge and the low-contrast edge.

Abstract: Methods and apparatus for the generation of interpolated frames of video data. In one embodiment, the interpolated frames of video data are generated by obtaining two or more frames of video data from a video sequence; determining frame errors for the obtained two or more frames from the video sequence, determining whether the frame errors exceed a threshold value; performing a multi-pass operation; performing a single-pass operation; performing frame blending; performing edge correction; and generating the interpolated frame of image data.

Abstract: Methods and systems for descriptor guided fast marching method based image analysis and associated systems are disclosed. A representative image processing method includes processing an image of a microelectronic device using a fast marching algorithm to obtain arrival time information for the image. The arrival time information is analyzed using a targeted feature descriptor to identify targeted features. The detection of defects is facilitated by segmenting the image. The segmented image can be analyzed to identify targeted features which are then labeled for inspection.

Abstract: A noise reduction unit reduces color noise of an input image to generate a noise-reduced image. A first generation unit generates a color difference signal from the noise-reduced image. A reduction processing unit generates hierarchical images including at least two or more reduced images from the noise-reduced image. A second generation unit generates coloring-suppressed color difference signals from the reduced images. A combining unit combines the color difference signal generated by the first generation unit with each of the color difference signals generated by the second generation unit. The second generation unit selects one of the color difference signals of the reduced images for each pixel based on magnitude of each of the color difference signals of the reduced images, to generate the coloring-suppressed color difference signals.

Abstract: Systems and methods for detecting crop marks depicted in digital images are disclosed. According to certain aspects, an electronic device may, in conjunction with enabling a user to customize a product design, facilitate a set of digital image processing functionalities to detect a set of crop marks depicted in a digital image associated with the product design. In particular, the systems and methods may detect a first set of lines depicted in a processed digital image indicative of content of interest extracted from the digital image, and detect a second set of lines depicted in a subtracted digital resulting from subtracting a set of contours from the processed digital image. The systems and methods may determine the set of crop marks from the first and second sets of lines.

Abstract: A character recognition method for a moving image includes extracting a region corresponding to a character string included in each frame of a moving image to be recognized. The method reads a character string from the extracted region and corrects the character string read from each frame, based on appearance rule information that specifies an appearance rule of a character string corresponding to the order of the frames, such that the appearance order of the read character string conforms to the appearance rule.

Abstract: A system for inputting and processing handwritten, multi-character text may comprise a handwriting recognition subsystem, a word completion subsystem, and an audio feedback system. The handwriting recognition system may be configured to capture a series of handwritten characters formed by a user and to convert the handwritten characters into a set of candidate partial text strings. The word completion subsystem may be configured to identify if a candidate partial text string constitutes a word segment and if so, generate one or both of (i) at least one clarifying word and (ii) at least one clarifying phrase that includes the clarifying word. The word segment may be an arbitrary string and not correspond to a valid complete word in a language associated with the system. The audio feedback subsystem may be configured to produce an audio representation of the word segment(s), the clarifying word(s), and the clarifying phrase(s).