Abstract: A method may include applying a first machine learning model trained to perform an open-set detection by at least identifying, based at least on image data indicative of one or more objects present in at least one environment in which one or more vehicles operate, at least one new class of objects. A dataset including a plurality of classes of objects may be updated to include the at least one new class of objects. In some cases, the dataset may be further updated to include a label associated with the at least one new class of objects. A second machine learning model may be trained, or in some cases updated, based at least on the updated dataset including the at least one new class of objects. Related systems and computer program products are also provided.

Abstract: An encoding method for a picture part encoded in a bitstream and reconstructed can involve refinement data encoded in the bitstream and determined based on at least one comparison of a version of a rate distortion cost computed using a data coding cost, a distortion between an original version of the picture part and a reconstructed picture part, and at least one other version involving one or more combinations of with or without a refinement by the refinement data, or a refinement either in or out of the decoding loop, or with or without a mapping or an inverse mapping.

Abstract: A method comprising signaling (102) in a data structure metadata representative of an energy consumption induced by encoding tools and/or features implemented to obtain a video stream representative of a sequence of pictures; and, associating (103) the data structure to the video stream for at least one subset of pictures representative of a period of the video stream; wherein at least one encoding tool or feature is associated to an information representative of an energy consumption depending on: a single reference picture size defined for the sequence of pictures; or, a total number of blocks of a given size in pictures signaled in the metadata for the at least one subset of pictures; or, a number of samples per square and rectangular blocks

Abstract: Video coding tools can be controlled by including syntax in a video bitstream that makes better use of video decoding resources. An encoder inserts syntax into a video bitstream to enable a decoder to parse the bitstream and easily control which tools combinations are enabled, which combinations are not permitted, and which tools are activated for various components in a multiple component bitstream, leading to potential parallelization of bitstream decoding.

Abstract: Motion vectors of a first reference frame are permitted to point to a plurality of further reference frames. A method of storing the motion vectors comprises, when a block of the first reference frame has two motion vectors (V2A, V2B) initially, selecting one of the two motion vectors, the non-selected motion vector not being stored. The selected motion vector may be scaled. This can reduce the motion vector memory size.

Abstract: Provided are methods, systems, and computer program products for generating an output map indicating a likelihood of individual elements of an image as corresponding to particular road elements, such as lane dividers, road dividers, and road boundaries. An example method may include applying a machine learning architecture to the image, which architecture includes a convolutional neural network and a sub-network capturing global context from feature maps generated by the convolutional neural network.

Abstract: Motion vectors of a first reference frame are permitted to point to a plurality of further reference frames. A method of storing the motion vectors comprises, when a block of the first reference frame has two motion vectors (V2A, V2B) initially, selecting one of the two motion vectors, the non-selected motion vector not being stored. The selected motion vector may be scaled. This can reduce the motion vector memory size.

Abstract: Motion vectors of a first reference frame are permitted to point to a plurality of further reference frames. A method of storing the motion vectors comprises, when a block of the first reference frame has two motion vectors (V2A, V2B) initially, selecting one of the two motion vectors, the non-selected motion vector not being stored. The selected motion vector may be scaled. This can reduce the motion vector memory size.

Abstract: An image is split into a plurality of blocks of various sizes and a subdivision level counter is associated to each of the blocks. The value of this subdivision level counter for a block is representative of the size of the block and is used to determine the quantization parameter for the block. The value is propagated for each subdivision and incremented according the type of the subdivision. When the image is split, an analysis is done according to the subdivision level counter, a maximal value of subdivision and the type of split, in order to determine the start of a new quantization group and when it is the case, the current position of the partition is propagated to the further split partitions to be stored with these partition and serve in the prediction process when decoding.

Abstract: A video coding system uses a value of a chroma quantization parameter table to quantize or de-quantize chroma values of a block of the video, the chroma quantization parameter table being directly coded in a set of parameters at the picture level or at the sequence level of the coded video stream. A corresponding encoding method, encoding device, decoding method, decoding device, and non-transitory readable medium are proposed.

Abstract: A method comprising: obtaining (601) a syntax element associated with video data, wherein said syntax element specifies that a cross-color component transform process in which at least one second color component is transformed based on at least one first color component different from the second color component is to be applied to the video data as a post processing process following a decoding process to be applied to the video data.

Abstract: A capacitive transducer is provided. The capacitive transducer includes a plate including a protruding center mass and a substrate with a center depression configured to accept the center mass. The capacitive transducer also includes a first electrode coupled to a non-horizontal edge surface of the center mass and a second electrode coupled to a non-horizontal edge surface of the center depression. The capacitive transducer further includes a third electrode coupled to a horizontal edge surface of the center mass and a fourth electrode coupled to a horizontal edge surface of the center depression. The plate is coupled to the substrate at least along an outer perimeter area of the plate and the substrate.

Abstract: Apparatuses and methods for encoding or decoding picture information that involve determining a scaling factor to scale a chroma prediction residual associated with chroma information included in the picture information; scaling the chroma prediction residual based on the scaling factor to provide a scaled chroma prediction residual; and encoding or decoding at least a portion of the picture information based on the scaled chroma prediction residual.

Abstract: A video coding system uses a value of a chroma quantization parameter table to quantize or de-quantize chroma values of a block of the video, the chroma quantization parameter table being directly coded in a set of parameters at the picture level or at the sequence level of the coded video stream. A corresponding encoding method, encoding device, decoding method, decoding device, and non-transitory readable medium are proposed.

Abstract: Provided are methods for detecting when a vehicle is encountering an out of operational design domain (ODD) scenario, which can include training a generative adversarial network (GAN) including a generator network and a discriminator network. The generator network may be trained to generate synthesized scenarios. The discriminator network may be trained to distinguish between true scenarios and the synthesized scenarios generated by the generator network. The trained discriminator network may be applied to detect when a vehicle encounters an out of operational design domain (ODD) scenario. Some methods described also include controlling the motion of the vehicle in response to an output of the trained discriminator network indicating that the vehicle is encountering the out of operational design domain (ODD) scenario. Systems and computer program products are also provided.

Abstract: Systems and methods are disclosed for identifying visibility of a physical space using visibility values and indications of uncertainty associated with the visibility values. One method can include capturing an image of a physical space, dividing pixels of the image into presumably well-lit and presumably not well-lit categories based on an intensity threshold, generating an evidential illumination map for the image based at least partly on a comparison between a value of each pixel to the intensity threshold, and projecting the evidential illumination map onto data representing a three-dimensional scan of the physical space. Evidential values can enable an autonomous vehicle to more safely navigate spaces using camera data by enabling programmatic determination of uncertainty for the camera data.

Abstract: Encoding or decoding picture information can involve determining a first scaling factor varying with a first granularity to scale a chroma prediction residual associated with chroma information included in the picture information; determining a second scaling factor varying with a second granularity finer than the first granularity; scaling the chroma prediction residual based on a combination of the first scaling factor and the second scaling factor to provide a scaled chroma prediction residual with the second granularity; and encoding or decoding at least a portion of the picture information based on the scaled chroma prediction residual.

Abstract: A method for video decoding pictures comprising: reconstructing (701) a picture at a first spatial resolution; and, obtaining (702) metadata associated to said picture representative of an information specifying that applying at least one in-loop filtering and/or at least one post-filtering on at least a portion of said reconstructed picture at a second spatial resolution different from the first resolution is allowed.

Abstract: Methods and apparatus for performing in-loop filtering in an encoder or a decoder provide regions which use a common set of filter parameters An index can be sent from an encoder to a decoder indicating which set of filter parameters is to be used for a particular region. The in-loop filter can be Sample Adaptive Offset, Adaptive Loop Filter, or any other such filter. An encoder classifies regions of a picture according to blocks using a common set of filter parameter. The classification can be in the form of a map. Filtering blocks use the common set of filter parameters for a region. A decoder parses a bitstream for a set of filter parameters and an index indicative of the filter parameter set for a region being decoded.

Abstract: Quantization matrix can be used to adjust quantization of transform coefficients at different frequencies. In one embodiment, a single fixed parametric model, such as a polynomial is used to represent a quantization matrix. Modulation of bit cost and complexity is achieved by specifying only the n first polynomial coefficients, the remaining ones being implicitly set to zero or other default values. One form of the single fixed polynomial is a fully developed polynomial in (x,y), where x,y indicate the coordinates of a given coefficient in a quantization matrix, with terms ordered by increasing exponent. Since higher exponents are the last ones, reducing the number of polynomial coefficients reduces the degree of the polynomial, hence its complexity. The polynomial coefficients can be symmetrical in x and y, and thus reducing the number of polynomial coefficients that need to be signaled in the bitstream.