A method for face recognition through facial expression normalization includes: fitting an input two-dimensional face image into a three-dimensional face model by using a three-dimensional face database; normalizing the three-dimensional face model into a neutral-expression three-dimensional face model by using a neutral-expression parameter learned from the three-dimensional face database; converting the neutral-expression three-dimensional face model into a neutral-expression two-dimensional face image; and recognizing the neutral-expression two-dimensional face image from a two-dimensional face database. Accordingly, face recognition may be performed with high reliability without a loss of information.

A system that includes a three-dimensional (3D) camera configured to capture 3D images of a dairy livestock. The system further includes a memory operable to store a thigh gap detection rule set and a processor operably coupled to the 3D camera and the memory. The processor is configured to obtain the 3D image, to identify one or more regions within the 3D image comprising depth values greater than a depth value threshold, and to apply the thigh gap detection rule set to the one or more regions to identify a thigh gap region. The processor is further configured to demarcate an access region within the thigh gap region. The processor is configured to reduce the width of the access region by shifting a first vertical edge and a second vertical edge of the access region and to determine position information for the first vertical edge and the second vertical edge.

An image of a physical environment is acquired that comprises a plurality of pixels, each pixel including a two-dimensional pixel location in the image plane and a depth value corresponding to a distance between a region of the physical environment and the image plane. For each pixel, the two dimensional pixel location and the depth value is converted into a corresponding three-dimensional point in the physical environment defined by three coordinate components, each of which has a value in physical units of measurement. A set of edge points is determined within the plurality of three-dimensional points based, at least in part, on the z coordinate component of the plurality of points and a distance map is generated comprising a matrix of cells. For each cell of the distance map, a distance value is assigned representing a distance between the cell and the closest edge point to that cell.

There is provided a method of determining a three-dimensional shape of an insert for insertion into an ear. The method includes receiving image data corresponding to a two-dimensional image of an ear, processing the image data to measure at least one biometric feature of the ear, the at least one biometric feature being indicative of a three-dimensional shape of at least part of the ear, and determining a three-dimensional shape of an insert for insertion into the ear by matching said at least one biometric feature with one of a plurality of pre-stored three-dimensional shapes. Each pre-stored three-dimensional shape corresponds to a respective ear.

Described is a robotic system for detecting obstacles reliably with their ranges by a combination of two-dimensional and three-dimensional sensing. In operation, the system receives an image from a monocular video and range depth data from a range sensor of a scene proximate a mobile platform. The image is segmented into multiple object regions of interest and time-to-contact (TTC) value are calculated by estimating motion field and operating on image intensities. A two-dimensional (2D) TTC map is then generated by estimating average TTC values over the multiple object regions of interest. A three-dimensional TTC map is then generated by fusing the range depth data with image. Finally, a range-fused TTC map is generated by averaging the 2D TTC map and the 3D TTC map.

Methods and systems are described which allow a classification of first and second objects in an X-ray projection image. A respective representation and localization of both objects are determined by applying the models to match the objects in the X-ray image and a spatial relation of the classified objects is obtained. Such methods and systems take advantage of artificial intelligence.

A vehicle service system having a means for acquiring images of a three-dimensional region of a vehicle wheel assembly tire tread surface. The vehicle service system is configured to process the acquired images to produce a collection of data points corresponding to the spatial position of surface points in the region from which tire tread wear characteristics are identified. The acquired images are further utilized to provide both a graphical and a numerical display to an operator, with the numerical display linked to specifically annotated or indexed points or windows within the graphical display, thereby enabling an operator to quickly identify specific focus points or regions on the tire surface which have been measured at the numerically identified tread depths.

According to one example for segmenting image data, image data comprising color pixel data, IR data, and depth data is received from a sensor. The image data is segmented into a first list of objects based on at least one computed feature of the image data. At least one object type is determined for at least one object in the first list of objects. The segmentation of the first list of objects is refined into a second list of objects based on the at least one object type. In an example, the second list of objects is output.

Some implementations of the disclosure are directed to techniques for facial reconstruction from a sparse set of facial markers. In one implementation, a method comprises: obtaining data comprising a captured facial performance of a subject with a plurality of facial markers; determining a three-dimensional (3D) bundle corresponding to each of the plurality of facial markers of the captured facial performance; using at least the determined 3D bundles to retrieve, from a facial dataset comprising a plurality of facial shapes of the subject, a local geometric shape corresponding to each of the plurality of the facial markers; and merging the retrieved local geometric shapes to create a facial reconstruction of the subject for the captured facial performance.

Systems and methods for automated stereology are provided. In some embodiments, an active deep learning approach may be utilized to allow for a faster and more efficient training of a deep learning model for stereology analysis. In other embodiments, existing deep learning models for stereology analysis may be re-tuned to develop greater accuracy for a given data set of interest, either with or without an active deep learning approach. A method can include: capturing a data set including a stack of images of a three-dimensional (3D) object; determining whether an existing deep learning model is appropriate for use on the stack of images (or for re-tuning); performing pre-processing on the data set; performing a training of a deep learning model; applying the deep learning model to obtain a confidence score for each label of the data set; reviewing, by a user, at least some labels in the active set to verify whether the label displays sufficient agreement with an expected result, and moving only those that display sufficient agreement to a training set; and performing a stereology analysis using the trained deep learning model.