Abstract: Systems/techniques that facilitate deep learning image analysis with increased modularity and reduced footprint are provided. In various embodiments, a system can access medical imaging data. In various aspects, the system can perform, via execution of a deep learning neural network, a plurality of inferencing tasks on the medical imaging data. In various instances, the deep learning neural network can comprise a common backbone in parallel with a plurality of task-specific backbones. In various cases, the plurality of task-specific backbones can respectively correspond to the plurality of inferencing tasks.

Abstract: According to one aspect, a method includes positioning a lidar on a platform, measuring at least a first data point using the lidar, measuring at last a second data point using the lidar, processing the at least first data point and the at least second data point to identify a first estimated location of the vertex, and determining whether the first estimated location of the vertex meets the specification, wherein when it is determined that the first estimated location of the vertex meets at least one specification associated with the lidar, the lidar is verified. The platform rotates about a first axis, wherein when the platform rotates about the first axis, the lidar rotates about the first axis. The data points measured with respect to a target, the target having at least a first edge, a second edge, and a vertex at which the first edge meets the second edge.

Abstract: According to one aspect, a method includes obtaining an assembly, the assembly including at least a first assembly sensor and a second assembly sensor of a plurality of assembly sensors, and calibrating the first assembly sensor and the second assembly sensor. The method also includes calibrating the first assembly sensor and the second assembly sensor together, wherein calibrating the first assembly sensor and second assembly sensor together includes placing the assembly in at least a first pose relative to a plurality of calibration targets. The assembly is coupled on a vehicle, and the plurality of assembly sensors are calibrated after coupling the assembly on the vehicle. Calibrating the plurality of assembly sensors after installing the assembly on the vehicle includes causing the vehicle to move relative to the plurality of calibration targets and using the plurality of assembly sensors to make measurements using the plurality of calibration targets.

Abstract: The invention provides a system and method that automatically generates a user interface (HMI) based on a selection of spreadsheet cells. The spreadsheet controls operations within the processor(s) of one or more vision system cameras. After selecting a range of cells in the spreadsheet, the user applies the system and method by pressing a button, or using a menu command that results in an automatically generated HMI with appropriate scaling of interface elements and a desired layout of such elements on the associated screen. Advantageously, the system and method essentially reduces the user's workflow to two steps, selecting spreadsheet cells and generating the HMI around them. The generated HMI runs in a web browser that can be instantiated on a user device, and communicates directly to a vision system processor(s). Data can pass directly between the user interface running in a web browser and the vision system processor(s).

Abstract: A method of implementing an artificial neural network, ANN, (100) comprises applying a splitting operation for each respective target portion (130b) of a target tensor (130a): i) determining a respective source portion (130a) of a source tensor (120a) required to produce that target portion (130b); ii) loading values from the determined source portion (130a, and not other values from the source tensor (120a), to a working memory (202a); iii) calculating the target portion (130b) using the source portion (130a) in the working memory (202a); iv) outputting the calculated target portion (130b) for storing in an output memory (202b).