In an embodiment, a pose estimation device obtains an image of an object, and generates a pose estimate of the object. The pose estimate includes a respective heatmap for each of a plurality of pose components of a pose of the object, and the respective heatmap for each of the pose components includes a respective uncertainty indication of an uncertainty of the pose component at each of one or more pixels of the image.

An apparatus includes a display control unit, a receiving unit, an adjusting unit, and a determination unit. The display control unit is configured to display an image showing a result of detection of a defect from a captured image of a structure on a display device. The receiving unit is configured to receive an operation to specify part of the displayed image as a first region and an operation to give an instruction to correct at least part of the detection data corresponding to the first region. The adjusting unit is configured to adjust a parameter to be applied to the first region according to the instruction. The determination unit is configured to determine one or more second regions to which the adjusted parameter is to be applied from a plurality of segmented regions of the image.

An image matching system for determining visual overlaps between images by using box embeddings is described herein. The system receives two images depicting a 3D surface with different camera poses. The system inputs the images (or a crop of each image) into a machine learning model that outputs a box encoding for the first image and a box encoding for the second image. A box encoding includes parameters defining a box in an embedding space. Then the system determines an asymmetric overlap factor that measures asymmetric surface overlaps between the first image and the second image based on the box encodings. The asymmetric overlap factor includes an enclosure factor indicating how much surface from the first image is visible in the second image and a concentration factor indicating how much surface from the second image is visible in the first image.

A method for locating a three-dimensional target with respect to a vehicle is disclosed including capturing an image of the target, and from a three-dimensional mesh of the target, and from an estimation of the pose of the target, determining a set of projecting edges of the mesh of the target in the pose. The step of determining the projecting edges of the mesh of the target includes positioning the mesh of the target according to the pose, projecting in two dimensions the mesh so positioned, scanning the projection of the mesh in a plurality of scanning rows and, for each scanning row: defining a set of segments, each segment corresponding to the intersection of a face of the mesh with the scanning row and being defined by its ends, analyzing the relative depths of the ends of the segments, the depth being the position along a third dimension orthogonal to the two dimensions of the projection, in order to select a set of end points of segments corresponding to projecting edges of the mesh.

A method and an apparatus for pattern recognition is provided in the present invention, applied to the field of artificial intelligence. The method includes: acquiring a two-dimensional image of a target object and a two-dimensional feature of the target object according to the two-dimensional image of the target object; and acquiring a three-dimensional image of the target object and a three-dimensional feature of the target object according to the three-dimensional image of the target object; identifying the target object according to the two-dimensional feature and the three-dimensional feature of the target object. The method can reduce restrictions on acquiring the image of the target object, for example, reduce the restrictions on the image of the target object in terms of postures, lighting, expressions, make-up and occlusion, thereby improving an accuracy of recognizing the target object and improving a recognition rate and reducing recognition time at the same time.

A training label image correction method includes performing a segmentation process on an input image (11) of training data (10) by a trained model (1) using the training data to create a determination label image (14), comparing labels of corresponding portions in the determination label image (14) and a training label image (12) with each other, and correcting label areas (13) included in the training label image based on label comparison results.

Systems and methods for detecting blind spots using a robotic apparatus are disclosed herein. According to at least one exemplary embodiment, a robot may utilize a plurality of virtual robots or representations to determine intersection points between extended measurements from the robot and virtual measurements from a respective one of the virtual robot or representation to determine blind spots. The robot may additionally consider locations of the blind spots while navigating a route to enhance safety, wherein the robot may perform an action to alert nearby humans upon navigating near a blind spot along the route.

Systems and methods described herein relate to detecting objects. One embodiment receives a plurality of three-dimensional (3D) data points from a plurality of light beams emitted by one or more sensors; identifies, among the plurality of 3D data points, a first set of inlier points that satisfy a first predetermined error condition with respect to a plane hypothesis and a first set of outlier points that fail to satisfy the first predetermined error condition; identifying, among the first set of inlier points, a second set of outlier points, the second set of outlier points failing to satisfy a second predetermined error condition in a range domain with respect to a plurality of line hypotheses corresponding, respectively, to the plurality of light beams; and detecting an object based, at least in part, on at least one of the first set of outlier points and the second set of outlier points.

Apparatus for diagnosing breast cancer, the apparatus comprising a controller having a set of instructions executable to: acquire a contrast enhanced region of interest (CE-ROI) in an X-ray image of a patient's breast, the X-ray image comprising X-ray pixels that indicate intensity of X-rays that passed through the breast to generate the image; determine a texture neighborhood for each of a plurality of X-ray pixels in the CE-ROI, the texture neighborhood for a given X-ray pixel of the plurality of X-ray pixels extending to a bounding pixel radius of BPR pixels from the given pixel; generate a texture feature vector (TF) having components based on the indications of intensity provided by a plurality of X-ray pixels in the CE-ROI that are located within the texture neighborhood; and use a classifier to classify the texture feature vector TF to determine whether the CE-ROI is malignant

A method for grading a precious gem. The method may include generating, by a signature generator, a signature of the precious gem; searching, out of multiple concept structures, for a matching concept structures that comprises at least one reference signature that matches the signature of the precious gem; wherein each concept structure comprises precious gem signatures of the same grade; wherein at least two concept structures differ from each other by grade; wherein each concept structure is generated by applying an unsupervised learning process and associating a grade with the cluster; and determining the grade of the precious gem based on a grade associated with a matching concept structure.