In an object detection device, a plurality of object detection units output a score indicating the probability that a predetermined object exists for each partial region set with respect to inputted image data. On the basis of the image data, a weight computation unit uses weight computation parameters to compute weights for each of the plurality of object detection units, the weights being used when the scores outputted by the plurality of object detection units are merged. A merging unit merges the scores outputted by the plurality of object detection units for each partial region according to the weights computed by the weight computation unit. A loss computation unit computes a difference between a ground truth label of the image data and the scores merged by the merging unit as a loss. Then, a parameter correction unit corrects the weight computation parameters so as to reduce the computed loss.

Systems and methods are described for image classification. An example method may comprise receiving an image comprising an object of interest and determining, based on a first convolution neural network, a first classification of the image. The first convolution neural network may be optimized for a first factor. The method may comprise determining, based on a second convolution neural network, a first classification of the image. The second convolution neural network may be optimized for a second factor. The method may comprise determining, based on the first classification and the second classification, a characteristic of the object of interest. The method may comprise providing the characteristic of the object of interest.

Described is a multiple-camera system and process for re-identifying an agent located in a materials handling facility based on anterior views of agents. An anterior view of a newly detected agent may be partitioned and color signatures generated for each partition. Likewise, stored anterior views of agents (candidate agents) that may potentially be the newly detected agent are partitioned and color signatures generated for each partition. Based on the color signatures, a similarity between the anterior view of the newly detected agent and the candidate agents is determined. The similarity may be used to either determine that the newly detected agent is one of the candidate agents or reduce the set of candidate agents that are considered during a manual review.

A plurality of recognition units respectively recognize image data using a learned model and output degrees of reliability corresponding to classes regarded as recognition targets by respective recognition units. A reliability generation unit generates degrees of reliability corresponding to a plurality of target classes based on the degrees of reliability output from the plurality of recognition units. A target model recognition unit recognizes the same image data as that recognized by the recognition units, by using a target model, and adjusts parameters of the target model in order to match the degrees of reliability corresponding to the target classes generated by a generation unit that outputs degrees of reliability corresponding to the target classes with the degrees of reliability corresponding to the target classes output from the target model recognition unit.

An apparatus for estimating a position in an automated valet parking system includes a front camera processor processing a front image of a vehicle, a surround view monitor (SVM) processor recognizing a short-distance lane and stop line by processing a surround view image of the vehicle, a map data unit storing a high definition map, and a controller downloading a map including an area set as a parking zone from the map data unit when the entry of the vehicle to a parking lot is identified and correcting a position measurement value of the vehicle by performing map matching based on results of the recognition and processing of the front camera processor and SVM processor and the parking lot map of the map data unit when an automated valet parking start position is recognized based on the recognized short-distance lane and stop line.

Provided are a hand pose estimation method, a device and a computer storage medium. The method may include: determining a classification logic map corresponding to each of a plurality of key-points, the plurality of key-points may represent key nodes of a skeleton of a target hand skeleton, a first key-point may be any one of the plurality of key-points; determining, based on a preset classification map and the classification logic map corresponding to the first key-point, co-ordinate information of the first key-point; and obtaining a pose estimation result of the target hand, in response to determining the co-ordinate information corresponding to each of the plurality of key-points.

In one embodiment, a method includes locating a plurality of part patches from an image, wherein each part patch comprises at least a portion of the image corresponding to a recognized human body portion or pose, and wherein each part patch is associated with a respective detection score larger than a threshold score, wherein the detection score is determined based on a comparison between the part patch with multiple training patches, generating a plurality of sets of feature data by processing each of the plurality of part patches with a plurality of convolutional neural networks, respectively, and determining whether a human attribute exists in the image based on the plurality of sets of feature data.

A driving assistance functionality for a motor vehicle when approaching a tollgate is disclosed. The method involves a step (S4) of calculating a probability of a tollgate being present based on at least two road context attributes that are determined from the motor vehicle and defining a road context ahead of said vehicle, said road context attributes being decorrelated from any concept of a tollgate. Examples of road context attributes: speed limit signs; marking lines on the ground; speed bumps or rumble strips on the ground; obstacles such as other vehicles; drivable space.

A method for processing breast tissue image data includes processing the image data to generate a set of image slices collectively depicting the patient's breast; for each image slice, applying one or more filters associated with a plurality of multi-level feature modules, each configured to represent and recognize an assigned characteristic or feature of a high-dimensional object; generating at each multi-level feature module a feature map depicting regions of the image slice having the assigned feature; combining the feature maps generated from the plurality of multi-level feature modules into a combined image object map indicating a probability that the high-dimensional object is present at a particular location of the image slice; and creating a 2D synthesized image identifying one or more high-dimensional objects based at least in part on object maps generated for a plurality of image slices.

The present invention relates to a method for acquiring an object information, the method comprising: obtaining an input image acquired by capturing a sea; obtaining a noise level of the input image; when the noise level indicates a noise lower than a predetermined level, acquiring an object information related to an obstacle included in the input image from the input image by using a first artificial neural network, and when the noise level indicates a noise higher than the predetermined level, obtaining a noise-reduced image of which the environmental noise is reduced from the input image by using a second artificial neural network, and acquiring an object information related to an obstacle included in the sea from the noise-reduced image by using the first artificial neural network.