An approach for improving determining a significant slice associated with a tumor from a volume of medical images is disclosed. The approach is based on the annotation of tumor range and the slice index in which the tumor appears to have the largest area. The approach infer a tumor growth classifier on sliding window of the volume slices and creates a discrete integral function out of the classifier predictions. The approach applies post processing on the discrete integral function which can include a smoothing function and a bias correction. The approach selects the slice index of maximum value from the post processing step.

An approach for improving determining a significant slice associated with a tumor from a volume of medical images is disclosed. The approach is based on the annotation of tumor range and the slice index in which the tumor appears to have the largest area. The approach infer a tumor growth classifier on sliding window of the volume slices and creates a discrete integral function out of the classifier predictions. The approach applies post processing on the discrete integral function which can include a smoothing function and a bias correction. The approach selects the slice index of maximum value from the post processing step.

The present application provides an action recognition method and apparatus and an electronic device. The method includes: if a target object is detected from a video frame, acquiring a plurality of images containing the target object, and optical-flow images of the plurality of images; extracting an object trajectory feature of the target object from the plurality of images, and extracting an optical-flow trajectory feature of the target object from the optical-flow images of the plurality of images; and according to the object trajectory feature and the optical-flow trajectory feature, recognizing a type of an action of the target object. Because it combines the time-feature information and the spatial-feature information of the target object, effectively increases the accuracy of the detection and recognition on the action type, and may take into consideration the detection efficiency at the same time, thereby improving the overall detection performance.

A door lock control device comprises: a door lock interface for communicating with a door lock; an imaging device; a controller for processing images to be acquired through the imaging device; and a storage medium. The controller determines each of first and second objects in the images as being either authorized or unauthorized depending on whether each of the first and second objects matches authentication data read from the storage medium, and controls the door lock through a door lock interface by referring to a distance between the first and second objects determined from the images when the first object is determined as being authorized and the second object is determined as being unauthorized.

This disclosure describes systems and methods for training a neural network with a training data set including data items labeled at different granularities. During training, each item within the training data set can be fed through the neural network. For items with labels of a higher granularity, weights of the network can be adjusted based on a comparison between the output of the network and the label of the item. For items with labels of a lower granularity, an output of the network can be fed through a conversion function that convers the output from the higher granularity to the lower granularity. The weights of the network can then be adjusted based on a comparison between the converted output and the label of the item.

Systems and methods are described for the fully automatic, template-free locating and extracting of a plurality of two-dimensional projections of particles in a micrograph image. A set of reference images is automatically assembled from a micrograph image by analyzing the image data in each of a plurality of partially overlapping windows and identifying a subset of windows with image data satisfying at least one statistic criterion compared to other windows. A normalized cross-correlation is then calculated between the image data in each reference image and the image data in each of a plurality of query image windows. Based on this cross-correlation analysis, a plurality of locations in the micrograph is automatically identified as containing a two-dimensional projection of a different instance of the particle of the first type. The two-dimensional projections identified in the micrograph are then used to determine the three-dimensional structure of the particle.

The present invention relates to a method for deriving a 3D data from image data comprising: receiving, from at least one camera, image data representing an environment; detecting, from the image data, at least one object within the environment; classifying the at least one detected object, wherein the method comprises, for each classified object of the classified at least one objects: determining a 2D skeleton of the classified object by implementing a neural network to identify features of the classified object in the image data corresponding to the classified object; and constructing a 3D skeleton for the classified object, comprising mapping the determined 2D skeleton to 3D.

The present invention relates to a method for deriving a 3D data from image data comprising: receiving, from at least one camera, image data representing an environment; detecting, from the image data, at least one object within the environment; classifying the at least one detected object, wherein the method comprises, for each classified object of the classified at least one objects: determining a 2D skeleton of the classified object by implementing a neural network to identify features of the classified object in the image data corresponding to the classified object; and constructing a 3D skeleton for the classified object, comprising mapping the determined 2D skeleton to 3D.

This application disclosures a method and system for detecting a centerline of a vessel. The method may include obtaining image data, wherein the image data may include vessel data; selecting two endpoints of the vessel based on the vessel data; transforming the image data to generate a transformed image based on at least one image transformation function; and determining a path of the centerline of the vessel connecting the first endpoint of the vessel and the second endpoint of the vessel to obtain the centerline of the vessel based on the transformed image. The two endpoints of the vessel may include a first endpoint of the vessel and a second endpoint of the vessel.

Computer implemented method for detecting a type of seat occupancy, comprising capturing, by means of an imaging device, an image of a seat, the image comprising depth data and intensity data, performing, by means of a processor device, a classifier algorithm on the captured image to determine a level of occupancy, wherein, if the determination indicates that the level of occupancy is above a predetermined threshold, the method comprises processing, by means of the processor device, the depth data with a convolutional neural network, to determine a type of occupation and wherein, if the determination indicates that the level of occupancy is below a predetermined threshold, the method comprises processing, by means of the processor device, the intensity data with a convolutional neural network to determine a type of occupation.