Abstract: Imaging objectives and inspection systems equipped with such imaging objectives are disclosed. The imaging objective may include a front objective configured to produce a diffraction limited intermediate image. The imaging objective may also include a relay configured to receive the intermediate image produced by the front objective. The relay may include three spherical mirrors positioned to deliver a projection of the intermediate image to a fixed image plane.

Abstract: A face recognition system of a residential environment identifies an individual present in the residential environment. The residential environment include a plurality of home devices and is associated with a group of different persons. The face recognition system identifies which person in the group is the individual and generate an operating instruction for a home device based on identity of the individual. For example, the face recognition system captures an image set of the individual's head and face and applies the image set to a machine learning model that is trained to distinguish between the different persons based on images of their heads and faces. The face recognition system can retrieve a personal profile of the identified individual, which includes settings of the home device for the identified individual. The face recognition system generates the operating instruction based on the personal profile.

Abstract: A method of evaluating image quality includes capturing a first digital image of a target object using a first digital camera of a smart device. A second digital image of a user of the smart device is captured using a second digital camera of the smart device. The second digital image includes an image of the user's facial expression. A quality index is generated for the first digital image by analyzing one or more features of the second digital image. Analyzing the second digital image includes determining the user's sentiment. The quality index is then associated with the first digital image.

Abstract: The present disclosure provides a method and device for counting pedestrians. The method comprises: reading a pedestrian depth image, and comparing the pedestrian depth image and a pre-acquired environmental mean image to acquire a foreground image; dividing the foreground image into a plurality of regions, detecting whether or not there is a step in an edge pixel point of each region, and detecting whether or not the region surface formed in each region coincides with the curved surface of the head top; determining that the currently detected region is the region of the head top when there is a step in the point and when the region surface coincides with the curved surface; and counting and outputting the number of pedestrians according to the region of the head top determined from the pedestrian depth image and the region of the head top determined from an adjacent pedestrian depth image.

Abstract: The embodiment of the present invention provides a human face recognition method and recognition system. The method includes that: a human face recognition request is acquired, and a statement is randomly generated according to the human face recognition request; audio data and video data returned by a user in response to the statement are acquired; corresponding voice information is acquired according to the audio data; corresponding lip movement information is acquired according to the video data; and when the lip movement information and the voice information satisfy a preset rule, the human face recognition request is permitted. By performing fit goodness matching between the lip movement information and voice information in a video for dynamic human face recognition, an attack by human face recognition with a real photo may be effectively avoided, and higher security is achieved.

Abstract: A camera, a storage device, and an image processing apparatus are included. The storage device stores a plurality of captured images that are captured by the camera. The image processing apparatus estimates a region of an object included in a captured image of a current time based on a captured image of a previous time, for each of multiple objects included in the captured image of the current time. The image processing apparatus evaluates whether to correct the estimated region of each object based on a combination region in which the estimated regions of the respective objects are combined, and on a difference region that is identified by a background image and the captured image of the current time. The image processing apparatus corrects the estimated regions of the respective objects based on a result of evaluation.

Abstract: An error compensation device includes at least one photographing module, a comparison module and a compensation module compensation module. The photographing module establish a space coordinate system relative to the object and comprises a first photographer and a second photographer which respectively photograph a first non-deformed laser speckle image on a first surface and a second non-deformed laser speckle image on a second surface, the first surface differs from the second surface by an azimuth. The comparison module respectively compares the first non-deformed laser speckle image before and after the displacement with the second non-deformed laser speckle image to calculate a displacement value between the two surfaces. The compensation module controls the movement of the object according to the displacement value. The error compensation utilizes the displacement of the non-deformed laser speckle image to obtain the absolute error of the object and compensate the error.

Abstract: A method for finding one or more candidate digital images being likely candidates for depicting a specific object comprising: receiving an object digital image depicting the specific object; determining, using a classification subnet of a convolutional neural network, a class for the specific object depicted in the object digital image; selecting, based on the determined class for the specific object depicted in the object digital image, a feature vector generating subnet from a plurality of feature vector generating subnets; determining, by the selected feature vector generating subnet, a feature vector of the specific object depicted in the object digital image; locating one or more candidate digital images being likely candidates for depicting the specific object depicted in the object digital image by comparing the determined feature vector and feature vectors registered in a database, wherein each registered feature vector is associated with a digital image.

Abstract: A system, including a processor in a vehicle and a memory, the memory including instructions to be executed by the processor to acquire a color image wherein a portion of the color image includes a color calibration pattern and operate the vehicle based on determining an illumination invariant image based on the color image and transform parameters ?v and ?r based on the color calibration pattern.

Abstract: Systems and techniques for training an auto-encoder on a single class are presented. In one example, a system trains an auto-encoder based on first data associated with a first class to generate a trained auto-encoder. The system also applies, using a multiplier, gain data indicative of a gain value to second data associated with the first class and third data associated with a second class to generate enhanced input data that represents a differentiation between the second data associated with the first class and the third data associated with the second class. An input enhancer comprises the trained auto-encoder and the multiplier. Furthermore, the system trains a convolutional neural network based on the enhanced input data to generate a trained convolutional neural network. The system also classifies the first class and the second class based on the input enhancer and the trained convolutional neural network.

Abstract: An object recognition device includes an acquisition unit configured to acquire a recognition target image that serves as an object to be recognized; a retrieval unit configured to search an image database storing a plurality of image data in association with tag information and retrieve a similar image that matches the recognition target image; and a recognition unit configured to recognize the object included in the recognition target image on the basis of tag information associated with a similar image obtained by the retrieval unit. The recognition may select the tag information that appears most frequently among the tag information associated with the similar images as a recognition result. The recognition unit may also compute a tag information reliability score from the similar image in the retrieval result and recognize an object taking into account said reliability score.

Abstract: The present application provides a shooting method and a shooting apparatus, relating to the field of wireless communications technologies. The method comprising: determining a subject according to sensor data of a target device and image information of a selectable subject; establishing a communication connection with at least one target device associated with the subject; and sending shooting data of the associated subject to the target device with which the connection has been established. According to the method and the apparatus in embodiments of the present application, a subject is determined according to sensor data of a target device and image information of a selectable subject; therefore, a signal source can be precisely matched with a physical device, that is, an object to which shooting data is transmitted can be determined accurately, to achieve precise transmission of the shooting data.

Abstract: The present invention relates to a method, system and non-transitory computer-readable recording medium for compensating brightness of ball images. According to one aspect of the invention, there is provided a method for compensating brightness of ball images. The method includes calculating photographed brightness distribution of a region corresponding to a ball in each of a plurality of images in which the ball is photographed, and compensating the photographed brightness distribution of the region corresponding to the ball in at least one of the plurality of images, with reference to reference brightness distribution.

Abstract: The present invention generally relates to a method forming a fingerprint image using a fingerprint sensing system, and specifically the formation of an improved fingerprint image by combining fingerprint image data from a plurality of subsequently acquired images of a finger. The invention also relates to a corresponding fingerprint sensing system and to a computer program product.

Abstract: One embodiment of the present invention sets forth a technique for incorporating a text-containing image into a digital. The technique includes analyzing a digital image to determine one or more text-allowed regions included in the digital image, and, for each of the one or more text-allowed regions, computing a numeric value based on a color contrast between pixels of a text-containing image and pixels of the text-allowed region, wherein the text-containing image is to be incorporated into one of the text-allowed regions included in the digital image. The technique further includes selecting a first text-allowed region based at least in part on the numeric value computed for each text-allowed region, and incorporating the text-containing image into the first text-allowed region included in the digital image.

Abstract: The present invention discloses a method of separating, identifying and characterizing cracks in rocks or other materials in three-dimensional (3D) space, which processes as follows to a volumetric digital image: 1) preprocessing digital image; 2) statistically analyzing basic information of the image including porosity, connectivity of each pore, and position, size, orientation and anisotropy of each pore-structure; 3) filtration: removing non-crack structures; 4) smoothening: smoothening and mending the image; 5) thinning: thinning the void structure into a thickness d (2 to 3 voxels recommended) in a direction with shortest extension; 6) separation: separating intersected cracks in a crack network by breaking the connections; 7) combination: combining those elongated cracks that are disconnected in the last step, and restoring cracks to the thickness before thinning, and eventually giving out the characterization of the cracks.

Abstract: Some embodiments of the invention provide a method for identifying geographic locations and for performing a fine-grained classification of elements detected in images captured from multiple different viewpoints or perspectives. In several embodiments, the method identifies the geographic locations by probabilistically combining predictions from the different viewpoints by warping their outputs to a common geographic coordinate frame. The method of certain embodiments performs the fine-grained classification based on image portions from several images associated with a particular geographic location, where the images are captured from different perspectives and/or zoom levels.

Abstract: Examples of the present disclosure describe systems and methods for scene reconstruction from bursts of image data. In an example, an image capture device may gather information from multiple positions within the scene. At each position, a burst of image data may be captured, such that other images within the burst may be used to identify common image features, anchor points, and geometry, in order to generate a scene reconstruction as observed from the position. Thus, as a result of capturing bursts from multiple positions in a scene, multiple burst reconstructions may be generated. Each burst may be oriented within the scene by identifying a key frame for each burst and using common image features and anchor points among the key frames to determine a camera position for each key frame. The burst reconstructions may then be combined into a unified reconstruction, thereby generating a high-quality reconstruction of the scene.

Abstract: Present invention provides a ranking convolutional neural network constructing method and an image processing method and apparatus thereof. The ranking convolutional neural network includes a ranking layer that is configured to sort an output of a previous layer of the ranking layer, generate an output of the ranking layer according to the sorted output, and output the output of the ranking layer to a next layer of the ranking layer. Using the ranking convolutional neural network enables obtaining an output feature corresponding to the input feature image through automatic learning. Compared with prior art methods that obtain features through manual calculation, the method of the present invention is superior in terms of reflecting the objective laws contained by the patterns of the actual scene. When applied to the field of image processing, the method can significantly improve the effect of image processing.

Abstract: The present disclosure is directed toward systems and methods for identifying and providing a user of a networking system with other users of the networking system that have a facial resemblance to the user (e.g., look-alike users). In particular, a facial resemblance system uses an image of a user to identify other users that have a facial resemblance to the user (e.g., look-alike users). The facial resemblance system can present look-alike users to the user based on a resemblance score and/or a commonality score. Further, the facial resemblance system can employ feedback from users to improve the accuracy of the facial resemblance system. Upon receiving one or more look-alike users, the user can then engage (e.g., communicate or socially connect) with one or more of the look-alike users.