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: A TDM MIMO FMCW radar comprises an array of physical receivers with a first spacing in a first direction and a plurality of physical transmitters arranged with a second spacing in said first direction. A virtual array signal of a range-Doppler bin relating to a scene with a moving object is processed by a phase compensation method, which introduces a phase ambiguity between the subarrays. A positive or negative spatial phase change rate with respect to the first direction is computed based on elements of the compensated virtual array signal corresponding to one subarray at a time. From this, based on the spacings, a spatial phase change between a pair of the subarrays is predicted. Next, a residual phase shift between said pair of subarrays is determined by comparing an actual phase shift of the compensated virtual array signal and the predicted spatial phase shift.

Abstract: A signed media bitstream comprises data units and signature units. Each signature unit is associated with one or more nearby data units and include at least one fingerprint derived from the associated data units and a digital signature of the at least one fingerprint. A storing method comprises: receiving a segment of the media bitstream; identifying N?2 instances of a repeating data unit in the received segment; pruning up to N?1 of the identified instances of the repeating data unit; and storing the received segment after pruning. A validation method comprises: receiving a segment of the media bitstream stored in accordance with the storing method; and validating a signature unit using a digital signature contained therein. Despite the pruning of the repeating data unit, the received associated data units can be successfully validated, either directly or indirectly, by means of different embodiments herein.

Abstract: A method of stabilizing bounding boxes for objects in a video stream comprises: receiving a video stream comprising a sequence of image frames; detecting an object in the image frames and generating a bounding box surrounding the object; measuring a noise level for the video stream; and temporally filtering the bounding box over a plurality of image frames based on the measured noise level, thereby stabilizing the bounding box in the video stream. The disclosure further relates to an image processing system.

Abstract: A method includes aligning a first image frame of a first stream of images from a first image sensor and a further first image frame of a second stream of images from a second image sensor according to a first alignment, aligning a second image frame of the first stream of images and a further second image frame of the second stream of images according to the first alignment, obtaining a first stitched image frame by blending the aligned first image frame and further first image frame in a first blending area according to a first blending function, obtaining second stitched image frame by blending the aligned second image frame and the further second image frame in a second blending area according to a second blending function, and evaluating the first alignment from a difference pattern calculated from pixel values of the first stitched frame and the second stitched frame.

Abstract: An arrangement for attachment of devices to a surface, such as a wall, comprises at least one device, and for each device, at least two mounting brackets. Each device has a rear side provided with a recess for each mounting bracket, wherein each mounting bracket has a male connector end and a female connector end, the male connector end being configured to be received by the female connector end of an adjoining mounting bracket associated to a neighbouring device. Each male connector end is provided with a through hole for receiving an attachment means, such as a screw. Each mounting bracket is insertable into its associated recess with an orientation in which the male connector end extends beyond a periphery of the device, and each mounting bracket is insertable into its associated recess with an orientation in which the female connector end is accessible from the periphery of the device.

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: The present disclosure relates to methods, systems and non-transitory computer-readable storage mediums for detecting an object of a first object type in a video sequence. A first algorithm is used to detect areas or objects in the scene as captured in the video stream that have an uncertain object type status. A second algorithm is used to provide a background model of the video sequence. For areas or objects having the uncertain object type status, the background model is used to check if the area or object is considered to be part of the background or the foreground in the video sequence. If the area or object is determined to belong to the foreground, the area or object is classified as the first object type. If the area or object is determined to not belong to the foreground, the area or object is not classified as the first object type.

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 and apparatus, including computer program products, implementing and using techniques for controlling the display of an overlay in a video. An overlay to be displayed in a video is defined. An overlay area is defined, which includes the overlay. At least one foreground object in the video is defined. A spatial overlap between the foreground object and the overlay area is determined. In response to determining that a size of the spatial overlap exceeds a first threshold, the entire overlay is stopped from being displayed within the overlay area.

Abstract: A method for classifying a detected object is disclosed. First and second object detectors detect first and second objects in first and second image frames, respectively, of a video sequence, and first and second probability scores respectively are calculated indicating a probability that the detected object belongs to a specific class. The second image frame is subsequent to the first image frame. The first object detector has a higher object detection precision and a longer processing time than the second object detector. The first and second object detections are performed in parallel. Reducing the first classification threshold or increasing the first probability score are performed if the first probability score is below a first classification threshold and the second probability score is above a second classification threshold. The first object is determined to belong to the specific class based on the probability scores and the classification thresholds.

Abstract: A device and a method for buffering a graphical overlay to be applied to an image is disclosed. A graphical overlay description specifying content, size, and position in the image of a graphical element of the graphical overlay is obtained, and the graphical overlay is divided into a plurality of sequential line fragments. For each line fragment it is determined, using the graphical overlay description, whether the line fragment overlaps a part of the graphical element. On condition that the line fragment overlaps a part of the graphical element, information representing the part of the graphical element is buffered in a buffer memory for the line fragment. On condition that the line fragment does not overlap any part of the graphical element, a run-length coding representing identical pixels is buffered in the buffer memory for the line fragment.

Abstract: A method (100), a device (600;700) and a system (800) for processing image data representing a scene for extracting features related to objects in the scene using a convolutional neural network are disclosed. Two or more portions of the image data representing a respective one of two or more portions of the scene are processed (S110), by means of a respective one of two or more circuitries, through a first number of layers of the convolutional neural network to form two or more outputs, wherein the two or more portions of the scene are partially overlapping. The two or more outputs are combined (S120) to form a combined output, and the combined output is processed (S130) through a second number of layers of the convolutional neural network by means of one of the two or more circuitries for extracting features related to objects in the scene.

Abstract: A method of performing better controlled switching between day mode and night mode imaging in a camera, where those illuminants contributing to the ambient light in day mode are considered when determining the visible light during night mode. Characteristic values of those illuminants are mixed with several levels of IR light to simulate the presence of an IR illuminator, and these characteristic values are compared to corresponding values derived from the color components of the ambient light in night mode in order to determine the IR proportion, and from that the amount of visible light.

Abstract: A method of providing an event-initiated video sequence, comprises a pre-event video sequence and a post-event video sequence. The pre-event image frames are stored in a non-processed format until such time that an event is taking place. By postponing the image processing of the pre-event video sequence, power savings are accomplished.