Abstract: A device, a non-transitory computer-readable storage medium, and a method of signing a metadata frame corresponding to an image frame of a sequence of image frames are disclosed. The metadata frame comprises metadata of one or more detected objects in the image frame, and the metadata of each detected object comprises coordinates defining a location in the image frame of the detected object. A digital signature is generated based on at least a subset of the metadata in the metadata frame, and additional metadata are added to the metadata frame. The additional metadata comprise the digital signature and predefined coordinates which define that the additional metadata comprise the digital signature. Furthermore, a device, a non-transitory computer-readable storage medium, and method of authenticating a digitally signed metadata frame corresponding to an image frame of a sequence of image frames are disclosed.

Abstract: A method and a device for setting a value of an object property in a sequence of metadata frames corresponding to a sequence of video frames is provided. An object is detected in a plurality of video frames. For a temporally first video frame a first value of an object property is determined and the object property is set to have the first value in a metadata frame corresponding to the first video frame. For each subsequent video frame, a subsequent value of the object property is determined and a deviation is calculated with respect to a value of the object property that was last set in a metadata frame. If the deviation exceeds a non-zero threshold, the object property is set to have the subsequent value in a metadata frame corresponding to the subsequent video frame. Otherwise no value of the object property is set in that metadata frame.

Abstract: The present disclosure generally relates to a method for weighting of features in a feature vector of an object detected in a video stream capturing a scene, comprising: determining a feature vector comprising a set of features for a detected object in the video stream; acquiring a reference feature vector of a reference model of the scene; and assigning a weight to at least one feature of the determined feature vector, wherein the weight for a feature of the determined feature vector depends on a deviation measure indicative of a degree of deviation of the feature from a corresponding feature of the acquired reference feature vector of the reference model.

Abstract: A plurality of feature vectors for moveable objects are received in a sequence of image frames captured during a first period. The feature vectors are received from a machine learning module trained to extract similar feature vectors in different image frames. An initial value is assigned to a first feature vector , or to a cluster of feature vectors identified in a second sequence of image frames preceding the first image frame. The indicator indicates whether vector is alive. For each subsequent image frame , the indicator is updated by: updating the value based on a similarity between the feature vector and the first feature vector or the cluster of feature vectors. If the value of the indicator indicates that the vector is alive, determining that there is a moveable object that is located in the captured scene at least the predetermined portion of the given period of time.

Abstract: There are provided encoding and decoding methods, and corresponding systems which are beneficial in connection to performing a search among regions of interest, ROIs, in encoded video data. In the encoded video data, there are independently decodable ROIs. These ROIs and the encoded video frames in which they are present are identified in metadata which is searched responsive to a search query. The encoded video data further embeds information which associates the ROIs with sets of coding units, CUs, that spatially overlap with the ROIs. In connection to independently decoding the ROIs found in the search, the embedded information is used to identify the sets of CUs to decode.

Abstract: Methods, systems and computer program products, for producing streams of image frames. Image frames in streaming video are segmented into background segments and instance segments. A background image frame containing the background segments is created. At least some of the instance segments are classified into movable objects of interest and movable objects of non-interest. During a background update time period, the background image frame is updated when a movable object of non-interest has moved to reveal a background area, to include the revealed background area in the background image frame. A foreground image containing the movable objects of interest is created. Blocks of pixels of the updated background and foreground image frames are encoded. A stream of encoded foreground image frames having a first frame rate is produced. A stream of encoded updated background image frames a second, lower frame rate is produced.

Abstract: A method of generating a segmentation outcome which indicates individual instances of one or more object classes for an image in a sequence of images is disclosed.

Abstract: A method of transmitting an image stream together with color metadata is disclosed. The method comprises capturing image frames of a first image stream, capturing image frames of a second image stream, using camera settings for prioritizing color capture, detecting an object in a first image frame of the first image stream and determining color information associated with the object based on the second image stream. The method further comprises adding the color information as color metadata to the first image stream, encoding the first image stream, and transmitting the encoded first image stream together with the color metadata. A corresponding device is also disclosed.

Abstract: A method and a device associate an object detection in a first frame with an object detection in a second frame using a convolutional neural (CNN) network trained to determine feature vectors such that object detections relating to separate objects are arranged in separate clusters. The CNN determines a reference set of feature vectors associated with the object detection in the first frame, and candidate sets of feature vectors associated with a respective one of identified areas corresponding to object detections in the second frame. A set of closest feature vectors is determined, and then measure of closeness to the reference set of feature vectors is determined for each candidate. A respective weight is determined for each object detection in the second frame. The object detection in the first frame is associated with one of the object detections in the second frame based on the assigned weights.

Abstract: Methods, systems and computer program products, for producing streams of image frames. Image frames in streaming video are segmented into background segments and instance segments. A background image frame containing the background segments is created. At least some of the instance segments are classified into movable objects of interest and movable objects of non-interest. During a background update time period, the background image frame is updated when a movable object of non-interest has moved to reveal a background area, to include the revealed background area in the background image frame. A foreground image containing the movable objects of interest is created. Blocks of pixels of the updated background and foreground image frames are encoded. A stream of encoded foreground image frames having a first frame rate is produced. A stream of encoded updated background image frames a second, lower frame rate is produced.

Abstract: A method and a controller for controlling a video processing unit to facilitate detection of newcomers in a first environment. The method comprises: capturing a thermal image of a human object in the first environment, the first environment being associated with a first climate; calculating, based on the thermal image, a thermal signature of a portion of the human object; determining that the human object has entered the first environment from a second environment when the thermal signature of the portion of the human object deviates from a predetermined thermal signature associated with the first environment, wherein the second environment is associated with a second, different, climate; and controlling the video processing unit to prioritize the human object over other human objects when processing video frames depicting the human object together with the other human objects.

Abstract: In a method for tracking an object in video-monitoring scenes, multiple feature vectors are extracted (722) and assembled (724) in point clouds, wherein a point cloud may be assembled for each tracklet, i.e. for each separate part of a track. In order to determine if different tracklets relate to the same or different objects the point clouds of each tracklet is compared (734). Based on the outcome of the comparison it is deduced if the first object and the second object may be considered to be the same object and, if so, the first object is associated (738) with the second object.

Abstract: A method of object re-identification in images of objects comprises providing a plurality of neural networks for object re-identification, wherein each of the plurality of neural networks is trained on image data with different sets of anatomical features, each set being represented by a reference vector; receiving a plurality of images of objects and an input vector representing anatomical features that are depicted in all of the plurality of images; comparing the input vector with the reference vectors for determining, according to a predefined condition, the most similar reference vector; and inputting image data of the plurality of objects to the neural network represented by the most similar reference vector for determining whether the plurality of objects have the same identity.

Abstract: A method, device and computer program product for training neural networks being adapted to process image data and output a vector of values forming a feature vector for the processed image data. The training is performed using feature vectors from a reference neural network as ground truth. A system of devices for tracking an object using feature vectors outputted by neural networks running on the devices.

Abstract: A method and system for tracking objects in a defined area compares image data of a detected object to profiles of persons that have entered the defined area to find the best match and connect the profile of the best match to the detected object. Identification profiles of persons that have been identified, by presenting their credentials, when entering the defined area are registered as candidates and are later matched with objects detected in the defined area. The system and method use the physical access control system of the defined area to reduce the number of candidates for the detected objects to the most likely candidates. The processing time and need for resources of the object tracking in the defined area are thereby reduced.

Abstract: A method and an image processing entity for applying a convolutional neural network to an image are disclosed. The image processing entity processes the image while using the convolutional kernel to render a feature map, whereby a second feature map size of the feature map is greater than a first feature map size of the feature maps with which the feature kernel was trained. Furthermore, the image processing entity repeatedly applies the feature kernel to the feature map in a stepwise manner, wherein the feature kernel was trained to identify the feature based on the feature maps of the first feature maps, wherein the feature kernel has the first feature map size.

Abstract: A method and system for action recognition in a video sequence is disclosed. The system comprises a camera configured to capture the video sequence and a server configured to perform action recognition. The camera comprises an object identifier that identifies an object of interest in an object image frame of the video sequence; an action candidate recognizer configured to apply a first action recognition algorithm to the object image frame to detect presence of an action candidate; an video extractor configured to produce action image frames of an action video sequence by extracting video data pertaining to a plurality of image frames from the video sequence; and a network interface configured to transfer the action video sequence to the server. The server comprises an action verifier configured to apply a second action recognition algorithm to the action video sequence to verify or reject that the action candidate is an action.

Abstract: In a method for tracking an object in video-monitoring scenes, multiple feature vectors are extracted (722) and assembled (724) in point clouds, wherein a point cloud may be assembled for each tracklet, i.e. for each separate part of a track. In order to determine if different tracklets relate to the same or different objects the point clouds of each tracklet is compared (734). Based on the outcome of the comparison it is deduced if the first object and the second object may be considered to be the same object and, if so, the first object is associated (738) with the second object.

Abstract: Methods and apparatus, including computer program products, for creating a quality annotated training data set of images for training a quality estimating neural network. A set of images depicting a same object is received. The images in the set of images have varying image quality. A probe image whose quality is to be estimated is selected from the set of images. A gallery of images is selected from the set of images. The gallery of images does not include the probe image. The probe image is compared to each image in the gallery and a match score is generated for each image comparison. Based on the match scores, a quality value is determined for the probe image. The probe image and its associated quality value are added to a quality annotated training data set for the neural network.

Abstract: A method for finding one or more candidate digital images being likely candidates for depicting a specific object comprising: receiving an object digital image depicting the specific object; determining, using a classification subnet of a convolutional neural network, a class for the specific object depicted in the object digital image; selecting, based on the determined class for the specific object depicted in the object digital image, a feature vector generating subnet from a plurality of feature vector generating subnets; determining, by the selected feature vector generating subnet, a feature vector of the specific object depicted in the object digital image; locating one or more candidate digital images being likely candidates for depicting the specific object depicted in the object digital image by comparing the determined feature vector and feature vectors registered in a database, wherein each registered feature vector is associated with a digital image.