Abstract: A computer-implemented method of modelling a common structure component, the method comprising, in a modelling computer system: receiving a plurality of captured frames, each frame comprising a set of 3D structure points, in which at least a portion of a common structure component is captured; computing a first reference position within at least one first frame of the plurality of frames; selectively extracting first 3D structure points of the first frame based on the first reference position computed for the first frame; computing a second reference position within a second frame of the plurality of frames; selectively extracting second 3D structure points of the second frame based on the second reference position computed for the second frame; and aggregating the first 3D structure points and the second 3D structure points, thereby generating an aggregate 3D model of the common structure component based on the first and second reference positions.

Abstract: Herein, a “perception statistical performance model” (PSPM) for modelling a perception slice of a runtime stack for an autonomous vehicle or other robotic system may be used e.g. for safety/performance testing. A PSPM is configured to receive a computed perception ground truth, and determine from the perception ground truth, based on a set of learned parameters, a probabilistic perception uncertainty distribution, the parameters learned from a set of actual perception outputs generated using the perception slice to be modelled. A simulated scenario is run based on a time series of such perception outputs (with modelled perception errors), but can also be re-run based on perception ground truths directly (without perception errors). This can, for example, be way to ascertain whether perception error was the cause of some unexpected decision within the planner, by determining whether such a decision is also taken in the simulated scenario when perception error is “switched off”.

Abstract: The present disclosure pertains generally to image feature extraction. Both transfer-learning and multi-task training approaches are considered. In one example, a machine learning model is trained to perform a geographic classification task of distinguishing between images captured in different geographic regions based on their visual content. In another example, a machine learning model is trained to perform an order recognition task of determining information about the order of an image sequence based on its visual content, where the order of the image sequence may be different than the order in which its constituent images were captured. A further example combines the two approaches. The knowledge gained by the ML model in learning one or more such tasks can be applied to a desired image recognition task, such as image segmentation, structure detection or image classification, e.g. with a pre-training/fine-tuning framework or a multi-task learning framework.

Abstract: A computer-implemented method of creating one or more annotated perception inputs, the method comprising, in an annotation computer system: receiving a plurality of captured frames, each frame comprising a set of 3D structure points, in which at least a portion of a common structure component is captured; computing a reference position within at least one reference frame of the plurality of frames; generating a 3D model for the common structure component by selectively extracting 3D structure points of the at least one reference frame based on the reference position within that frame; determining an aligned model position for the 3D model within a target frame of the plurality of frames based on one or more manual alignment inputs received in respect of the target frame at a user interface whilst rendering the 3D model for manually aligning the 3D model with the common structure component in the target frame; and storing annotation data of the aligned model position in computer storage, in association with at

Abstract: A method of annotating road images, the method comprising implementing, at an image processing system, the following steps: receiving a time sequence of two dimensional images as captured by an image capture device of travelling vehicle; processing the images to reconstruct, in three-dimensional space, a path travelled by the vehicle; using the reconstructed vehicle path to determine expected road structure extending along the reconstructed vehicle path; and generating road annotation data for marking at least one of the images with an expected road structure location, by performing a geometric projection of the expected road structure in three-dimensional space onto a two-dimensional plane of that image.

Abstract: A computer-implemented method of training a depth uncertainty estimator comprises receiving, at a training computer system, a set of training examples, each training example comprising (i) a stereo image pair and (ii) an estimated disparity map computed from at least one image of the stereo image pair by a depth estimator. The training computer system executes a training process to learn one or more uncertainty estimation parameters of a perturbation function, the uncertainty estimation parameters for estimating uncertainty in disparity maps computed by the depth estimator. The training process is performed by sampling a likelihood function based on the training examples and the perturbation function, thereby obtaining a set of sampled values for learning the one or more uncertainty estimation parameters.

Abstract: A computer-implemented method of perceiving structure in an environment comprises steps of: receiving at least one structure observation input pertaining to the environment; processing the at least one structure observation input in a perception pipeline to compute a perception output; determining one or more uncertainty source inputs pertaining to the structure observation input; and determining for the perception output an associated uncertainty estimate by applying, to the one or more uncertainty source inputs, an uncertainty estimation function learned from statistical analysis of historical perception outputs.

Abstract: An image processing technique is presented using a hierarchical image model. The technique may be used as a precursor to subsequent image processing, to fix detected images in a post processing stage or as a segmentation or classification stage. The techniques may also be applied to super resolution. In a first layer of the hierarchical image model, each observed pixel of the image has a representation allocated to one or more input node. A set of the input nodes are assigned to a hidden node of a second layer, and a duplicate set of input nodes of the first layer is assigned to a duplicate of the hidden node in the second layer. In this way, a dense latent tree is formed in which a subtree is duplicated. Variables are assigned to the input nodes, the hidden node and the duplicate nodes and recurringly modified to process the image. Belief propagation messages may be utilised to recursively modify the variables. An image processing system using the method is described.

Abstract: A computer-implemented method of predicting an external actor trajectory comprises receiving, at a computer, sensor inputs for detecting and tracking an external actor; applying object tracking to the sensor inputs, in order track the external actor, and thereby determine an observed trace of the external actor over a time interval; determining a set of available goals for the external actor; for each of the available goals, determining an expected trajectory model; and comparing the observed trace of the external actor with the expected trajectory model for each of the available goals, to determine a likelihood of that goal.

Abstract: An in-stream disparity processing system comprises a delay block having an input for receiving an input stream of disparity cost vectors, and configured to provide a delayed stream of disparity cost vectors at an output of the delay block, by delaying the input stream by a predetermined amount; and a processing block having a cost input connected to receive the delayed stream of disparity cost vectors and a smoothing input connected to receive the input stream of disparity cost vectors, and configured to apply cost smoothing to the delayed stream based on the input stream, so as to generate, at an output of the processing block, a stream of reverse pass disparity cost vectors.

Abstract: The invention provides a computer-implemented method of planning a path for a mobile robot such as an autonomous vehicle in the presence of K obstacles. The method uses, for each of the K obstacles, a shape Bk and a density function pk(x) representing the probabilistic position of the obstacle. The method repeats the following steps for at least two different paths A:—choosing a path A, where A is the swept area of the robot within a given time interval; and—calculating based on the density function of each obstacle and the swept path an upper bound on the total probability of at least one collision FD between the robot and the K obstacles. This allows a number of candidate paths to be ranked for safety. By precomputing factors of the computational steps over K obstacles, the computation per path is O(N), and not O(NK). A safety threshold can be used to filter out paths below that threshold.

Abstract: A perception model is trained to classify inputs in relation to a discrete set of leaf node classes. A hierarchical classification tree encodes hierarchical relationships between the leaf node classes. A training loss function is dependent on a classification score for a given training input a its ground truth leaf node class of the training input, but also classification scores for at least some others of the leaf node classes, with the classification scores of the other leaf node classes weighted in dependence on their hierarchical relationship to the ground truth leaf node class within the hierarchical classification tree.

Abstract: The present disclosure pertains generally to image feature extraction. Both transfer-learning and multi-task training approaches are considered. In one example, a machine learning model is trained to perform a geographic classification task of distinguishing between images captured in different geographic regions based on their visual content. In another example, a machine learning model is trained to perform an order recognition task of determining information about the order of an image sequence based on its visual content, where the order of the image sequence may be different than the order in which its constituent images were captured. A further example combines the two approaches. The knowledge gained by the ML model in learning one or more such tasks can be applied to a desired image recognition task, such as image segmentation, structure detection or image classification, e.g. with a pre-training/fine-tuning framework or a multi-task learning framework.

Abstract: A method of annotating road images, the method comprising implementing, at an image processing system, the following steps: receiving a time sequence of two dimensional images as captured by an image capture device of travelling vehicle; processing the images to reconstruct, in three-dimensional space, a path travelled by the vehicle; using the reconstructed vehicle path to determine expected road structure extending along the reconstructed vehicle path; and generating road annotation data for marking at least one of the images with an expected road structure location, by performing a geometric projection of the expected road structure in three-dimensional space onto a two-dimensional plane of that image.