Abstract: A method for accessing and processing data by an accelerator apparatus includes retrieving at least a part of input data to be processed by the accelerator apparatus, segmenting the input data into a plurality of adjacent input tiles, the input tiles having a pre-determined size, storing at least one of the plurality of adjacent input tiles in a data buffer of the accelerator apparatus, accessing data elements of the at least one of the plurality of adjacent input tiles stored in the data buffer by positioning an input window over the data elements stored in the data buffer, to generate a plurality of input areas, wherein the input window is adjustable according to a set of parameters, and sequentially processing at least one of the plurality of input areas to at least partially process the at least one of the plurality of adjacent input tiles by the accelerator apparatus.

Abstract: The invention is a method for object segmentation in an image, comprising the steps of inputting the image to a trained machine learning system, and reconstructing the segmentation contour of the object. The method is characterized by comprising the steps of estimating, by the trained machine learning system, a representation of a segmentation contour of an object in the image, wherein the segmentation contour is a closed two-dimensional parametric curve, each point of which is defined by two coordinate components, wherein both coordinate components are parametrized, and wherein the reconstruction of the segmentation contour of the object is carried out from the estimated representation of the segmentation contour. The invention further relates to a data processing system, a computer program product and a computer readable medium carrying out the above method.

Abstract: A method for accessing and processing data by an accelerator apparatus includes retrieving at least a part of input data to be processed by the accelerator apparatus, segmenting the input data into a plurality of adjacent input tiles, the input tiles having a pre-determined size, storing at least one of the plurality of adjacent input tiles in a data buffer of the accelerator apparatus, accessing data elements of the at least one of the plurality of adjacent input tiles stored in the data buffer by positioning an input window over the data elements stored in the data buffer, to generate a plurality of input areas, wherein the input window is adjustable according to a set of parameters, and sequentially processing at least one of the plurality of input areas to at least partially process the at least one of the plurality of adjacent input tiles by the accelerator apparatus.

Abstract: Methods, systems, and computer program products for lane detection. An image processing module is trained by machine learning, and used to generate correspondence mapping data based on an image pair of a first and second images. The correspondence mapping data defines correspondence between a first lane boundary group of the first image and a second lane boundary group of the second image. An image space data block of image space detection pairs is then generated based on the correspondence mapping data, and a three-dimensional lane detection data block generated using triangulation based on first and second parts of the image space data block corresponding to respective first and second members of the image space detection pairs.

Abstract: A method for generating a mask based on a camera stream includes obtaining an image of the camera stream, generating a difference image, by subtracting two consecutive images from each other, one being the obtained image, providing a previous accumulated difference image and generating an accumulated difference image, generating an edge image by detecting edges in the obtained image, providing a previous accumulated edge image and generating an accumulated edge image, generating a combined image by combining the accumulated edge image and the accumulated difference image, defining a first threshold pixel value for the combined image, and generating the mask by including in the mask all pixels of the combined image, which have a same relation to the first threshold pixel value as that of a center pixel of the combined image.

Abstract: Sensor data is recorded during testing of a control unit of an autonomous car in an autonomous driving mode. In response to being triggered by a disengagement event, recorded sensor data from an analysis time window prior to the disengagement event is logged. The logged sensor data is processed in a simulator using the control unit to control a virtual autonomous car to identify the inadequately operating component of the control unit. The component may be modified and retested until the disengagement event is avoided in the simulator. The control unit with the modified component may then be used to record new sensor data during testing of the control unit in the autonomous driving mode, and the simulation repeated on the new data.

Abstract: The invention is a method and an apparatus for generating a displacement map of a first input dataset and a second input dataset of an input dataset pair. The apparatus comprises a neural network based feature extractor (25) for processing the first input dataset and the second input dataset so as to generate a feature map hierarchy (50) comprising a base pair of feature maps (20e, 30e) and a refinement pair of feature maps (20b, 30b, 20c, 30c, 20d, 30d), a displacement unit comprising a first comparator unit for obtaining an initial displacement map using the base pair of feature maps (20e, 30e); and a displacement refinement unit for obtain an updated displacement map for the refinement pair of feature maps (20d, 30d).

Abstract: The invention provides an autonomous vehicle capable of driving independently through a path of intense traffic and transporting objects or people, even on rough surfaces, whilst ensuring safety of the vehicle and general road safety. The autonomous vehicle includes at least one variable pitch camera for producing images to be used for a computer vision to control the autonomous vehicle. The invention facilitates changing the pitch of the variable pitch camera to maximise camera image clarity and/or resolution as the autonomous vehicle moves. The images can be used for lane detection, pedestrian detection, three-dimensional (3D) reconstruction of an environment and/or pothole detection. With the invention, at least one image from the variable pitch camera is used in controlling the movement and/or the trajectory of the autonomous vehicle.

Abstract: An electronic route navigation method is used in navigating an autonomous vehicle in a road network. In an implementation of the method a map is stored on at least one computer, and is sectioned into quads by a processor. The smallest size quads are grouped as base level quads. Extended area quads are defined to contain the area of a plurality of base level quads. Thereafter, a route along a road is calculated from a start point to a destination point. Further extended area quads on the map that do not contain the start point or the destination point are used for cross routing. In said cross routing only the road routes crossing the extended area quad leading from the start point to destination point are calculated.