Abstract: A depth map is generated from at least a first and a second image. A plurality of reference pixels are selected in the first image. A cost function is used to associate each reference pixel with a respective pixel in the second image. A masking operation is used to identify a subset of pixels in a block of pixels surrounding a reference pixel and the cost function is based on the identified subset of pixels. A disparity between each reference pixel and the respective pixel in said second image is determined, and a depth value is determined for each reference pixel as a function of the respective disparity. A depth map is generated based on the determined depth values.

Abstract: A system and method for computer vision based tracking of a human form may include detecting a shape of an object in an image of a space and determining the probability of object having a human form shape based on movement of the object. If the probability of the object of being of a human form is above a predetermined threshold the object is tracked and if the probability is below the threshold then the tracking is terminated. Occupancy in the space may be determined based on the tracking of the object.

Abstract: A 5D printer, which additively manufactures an object, includes an extruder that can move linearly along three orthogonal axes and rotationally around at least one of the axes with respect to the object while depositing a material. A gantry is movable along X, Y and Z axes, and a trunnion table movable about A and B axes is mounted on the gantry. A platen is mounted on the trunnion table, and the extruder deposits the material on the platen while moving the gantry and trunnion table. A model of the object is analyzed to produce a stress tensor for the object, and the depositing is according to the stress tensor.

Abstract: A structure for determining a plane in a depth image includes dividing a portion of a depth image into a plurality of areas, fitting a two-dimensional line to depth points in each of the plurality of areas, and combining two or more of the plurality of two-dimensional lines to form a three-dimensional plane estimate.

Abstract: A moving object detection method according to an embodiment includes evaluating an average value, instead of a median value, of pixel values at the same pixel positions across multiple superposed images. A threshold value obtained by multiplying the average value by a predetermined number is employed to distinguish space debris from an object which emits light at a higher luminance than the space debris. In order to detect the space debris accurately, an average value of the pixel values that are equal to or less than the threshold is employed as an evaluation value for the space debris detection.

Abstract: A passenger counting system that can count correctly the number of persons present in a vehicle, including persons sitting on the back seat, is provided. The passenger counting system includes an image capturing device 11 which is arranged outside the vehicle and which captures an image of the vehicle from the lateral direction of the vehicle, an image acquisition unit 31 that controls the image capturing device 11 and acquires an image in which the inside of the vehicle is captured as an input image, a profile detection unit 32 that detects a profile of a person from a partial image obtained from the input image based on an image feature amount calculated by using the partial image and outputs a detection result, and a passenger number determination unit 33 that determines the number of persons present in the vehicle based on the detection result.

Abstract: A portable device including a display unit, a control button and a controller. The display unit displays an execution screen of an application of the portable device. The control button receives a control input including at least one of a fingerprint input corresponding to a security authentication for the application and a push input corresponding to a cancel function. The controller displays a security authentication screen of the application, for which the fingerprint input is needed, receives a first control input including a first fingerprint input and a first push input when the security authentication screen of the application is displayed, and recognizes the first fingerprint input of the first control input by activating a fingerprint recognition function when the first fingerprint input is detected for more than a first threshold time and the first push input of the first control input is not detected.

Abstract: Systems, apparatuses, and methods are provided for refining building alignment in an aerial image. At least one candidate shifting vector and matching score value are determined for a local building. At least one dominant shifting vector is determined for at least one random group of neighboring buildings of the local building. At least one optimized matching score is calculated using the at least one candidate shifting vector for the local building and the at least one dominant shifting vector for the at least one random group of the neighboring buildings. A final shifting vector for the local building is found using the at least one optimized matching score.

Abstract: Images captured by multi-camera arrays with overlap regions can be stitched together using image stitching operations. An image stitching operation can be selected for use in stitching images based on a number of factors. An image stitching operation can be selected based on a view window location of a user viewing the images to be stitched together. An image stitching operation can also be selected based on a type, priority, or depth of image features located within an overlap region. Finally, an image stitching operation can be selected based on a likelihood that a particular image stitching operation will produce visible artifacts. Once a stitching operation is selected, the images corresponding to the overlap region can be stitched using the stitching operation, and the stitched image can be stored for subsequent access.

Abstract: Objects are tracked within images. According to an example embodiment, video data is processed for tracking one or more objects. A computer circuit executes instructions to process a series of ordered video frames in a set of video data as follows. A region of interest is defined in an initial one of the video frames, and a particle mesh is formed from a set of feature points for an object in the defined region. The particle mesh is propagated to an adjacent video frame using motion vectors of points in the mesh to form a coarse boundary at the adjacent frame. Occlusion and scene boundaries are detected in the adjacent video frame, and the detected boundaries are used to set a boundary and a number of feature points within the boundary in the adjacent video frame to maintain mesh coherence and track the object in the subsequent video frames.

Abstract: A method and related apparatus (116) for iterative reconstruction of a volume. A regularization parameter (?) of an iterative update function is spatially adapted and normalized in respect to a region of interest ROI in the volume. The method allows achieving essentially uniform noise distribution across the reconstructed volume.

Abstract: A method of estimating absolute orientation of a vehicle utilizing a camera and GPS. At least three images are selected from concurrent images taken above an essentially horizontal ground at a distance from each other. Key points are identified in each selected image. Absolute positions of the images are obtained from the GPS defining the image positions when a respective image was taken. Key points between selected images are matched to obtain an estimate of the absolute orientation.

Abstract: In an example embodiment, a method, apparatus and computer program product are provided. The method includes facilitating receipt of a plenoptic image associated with a scene, the plenoptic image including plenoptic micro-images and being captured by a focused plenoptic camera. The method includes generating plenoptic vectors for the plenoptic micro-images of the plenoptic image, where an individual plenoptic vector is generated for an individual plenoptic micro-image. The method includes assigning disparities for the plenoptic micro-images of the plenoptic image.

Abstract: Automated analysis of video data for determination of human behavior includes segmenting a video stream into a plurality of discrete individual frame image primitives which are combined into a visual event that may encompass an activity of concern as a function of a hypothesis. The visual event is optimized by setting a binary variable to true or false as a function of one or more constraints. The visual event is processed in view of associated non-video transaction data and the binary variable by associating the visual event with a logged transaction if associable, issuing an alert if the binary variable is true and the visual event is not associable with the logged transaction, and dropping the visual event if the binary variable is false and the visual event is not associable.

Abstract: A method and apparatus for three dimensional viewing of images is presented. Left eye viewing point (LEVP) imagery is passed through a left eye filter to obtain a filtered LEVP imagery. Right eye viewing point (REVP) imagery is passed through a right eye filter to obtain a filtered REVP imagery. The filtered LEVP imagery is projected on a display and the filtered REVP imagery is projected on the display. A user wearing polarized glasses is able to view the filtered LEVP imagery and the filtered REVP imagery as three-dimensional imagery on the display. The three dimensional imagery can be augmented.

Abstract: There is provided an image processing apparatus including a tone mapper that losslessly tone maps image data in a floating-point representation, a base layer generator that generates base layer image data by reducing a bit depth of the image data tone mapped by the tone mapper, and an enhancement layer generator that generates enhancement layer data by using the image data tone mapped by the tone mapper and the base layer image data generated by the base layer generator.

Abstract: Among other things, systems and techniques are provided for detecting defects on a wafer based upon non-correctable error data yielded from a scan of the wafer to determine a topology of the wafer. The non-correctable error data is reconstructed to generate a non-correctable error image map, which is transformed to generate a projection. In some embodiments, the non-correctable error image map is transformed via a feature extraction transform such as a Hough transform or a Radon transform. In some embodiments, the projection is compared to a set of rules to identify a signature in the non-correctable error image map indicative of a defect.

Abstract: Embodiments provide a method for processing an input image having a plurality of pixels, wherein each pixel has a pixel position. The method may include determining, for each pixel position, a vector based on the input image; and determining, for each pixel position, a detail value for the pixel position based on the determined vectors at a plurality of neighboring pixel positions within a predetermined neighboring block of the pixel position.

Abstract: The inventions herein relate generally to improvements in photogrammetry and devices suitable for obtaining such improvements. Some embodiments use only a single passive image-capture device to obtain overlapping 2D images, where such images at least partially overlap with regard to at least one object of interest in a scene. Such images can be processed using methods incorporating structure from motion algorithms. Accurate 3D digital representations of the at least one object of interest can be obtained. Substantially accurate measurements and other useful information regarding the at least one object of interest are obtainable from the methodology herein.

Abstract: Imaging data regarding an interior portion of a storage vessel may be obtained from an imaging device (e.g., a digital camera or depth sensor) and interpreted in order to identify an available volume within the storage vessel. Where the imaging data includes an image of an interior of the storage vessel, an area of one or more contiguous spaces of a rear face of the storage vessel may be identified from the image and multiplied by a depth of the storage vessel in order to estimate the available volume. Where the imaging data includes a depth profile of the interior, the depth profile may be processed in order to estimate the available volume. Once an available volume has been estimated, the capacity of the available volume to accommodate one or more additional items may be determined, and information regarding the capacity may be provided to a user.