Abstract: A method for determining a background in 3D space includes: receiving, via a sensor, a plurality of data points for the 3D space; dividing the 3D space into a plurality of rays extending from the sensor; defining a plurality of peaks for each ray; and defining a catchment region for each peak for each ray. Each data point is enclosed by a corresponding ray. Each peak is located at a peak distance from the sensor and comprises a peak height. The catchment region for each peak includes a catchment distance from the corresponding peak towards and away from the sensor. For each data point from the plurality of data points, the method includes: determining the corresponding ray enclosing the data point; determining a containing peak for which the data point lies within the catchment region; and incrementing the peak height of the containing peak by a peak increment value.

Abstract: A method for determining a background includes: receiving, via a sensor, a plurality of data points for a 3D space; dividing the 3D space into a plurality of rays extending from the sensor; defining a plurality of peaks for each ray; defining a catchment region for each peak for each ray and including a catchment distance from the corresponding peak towards and away from the sensor; and updating the catchment distance of the catchment region. Each data point is enclosed by a corresponding ray. Each peak is located at a peak distance from the sensor and includes a peak height. For each data point from the plurality of data points, the method includes: determining the corresponding ray enclosing the data point; determining a containing peak for which the data point lies within the catchment region; and incrementing the peak height of the containing peak by a peak increment value.

Abstract: A method for updating a cartesian representation includes receiving a sensor POV data based on a sensor data generated by at least one sensor for a plurality of view cones and generating a polar feature vector based on the sensor POV data. The polar feature vector includes a plurality of cone vectors corresponding to the plurality of view cones. The method further includes mapping at least one cell from a plurality of cells of the cartesian representation with at least one cone vector from the plurality of cone vectors, receiving at least one offset vector corresponding to the at least one cell, concatenating at least the at least one cone vector and the at least one offset vector to generate at least one transformed tensor for the at least one cell, and generating at least one learned output based on the at least one transformed tensor.

Abstract: A convolutional neural network system comprising at least one sensor, a convolutional neural network, and a weight generation network is provided. The convolutional neural network includes a plurality of convolution layers including corresponding one or more convolution kernels to generate one or more feature maps. The weight generation network includes a plurality of weight generators. One or more of the weight generators are configured to dynamically generate one or more weights for the one or more convolution kernels of one or more of the convolution layers at a run-time based on a vector. The vector is based on one or more weight indicators indicative of the one or more weights generated by a previous weight generator.