This document discloses system, method, and computer program product embodiments for controlling a vehicle. For example, the method includes: receiving an observation probability distribution function associated with a target object that was detected by sensor(s) of an autonomous vehicle (AV); identifying a target attribute associated with the target object; detecting a target attribute value associated with the target attribute; and issuing vehicle control instruction(s) that cause AV to adjust driving operation(s) using a future behavior of the target object predicted based on an attribute probability distribution function that defines a probability that the target attribute is actually present for the target object based on probability distribution function(s), wherein the attribute probability distribution function comprises: a probability value associated with the target attribute being present for the target object; and a probability value associated with the target attribute not being present for the target object.

This disclosure relates to improved techniques for performing human parsing functions using neural network architectures. The neural network architecture can model human objects in images using a hierarchal graph of interconnected nodes that correspond to anatomical features at various levels. Multi-level inference information can be generated for each of the nodes using separate inference processes. The multi-level inference information for each node can be combined or fused to generate final predictions for each of the nodes. Parsing results may be generated based on the final predictions.

A method for performing a computer-aided diagnosis (CAD) for universal lesion detection includes: receiving a medical image; processing the medical image to predict lesion proposals and generating cropped feature maps corresponding to the lesion proposals; for each lesion proposal, applying a plurality of lesion detection classifiers to generate a plurality of lesion detection scores, the plurality of lesion detection classifiers including a whole-body classifier and one or more organ-specific classifiers; for each lesion proposal, applying an organ-gating classifier to generate a plurality of weighting coefficients corresponding to the plurality of lesion detection classifiers; and for each lesion proposal, performing weight gating on the plurality of lesion detection scores with the plurality of weighting coefficients to generate a comprehensive lesion detection score.

A vison-based self-checkout terminal is provided. Purchased items are placed on a base and multiple cameras take multiple images of each item placed on the base. A location for each item placed on the base is determined along with a depth and the dimensions of each item at its given location on the base. Each item's images are then cropped, and item recognition is performed for each item on that item's cropped images with that item's corresponding depth and dimension attributes. An item identifier for each item is obtained along with a corresponding price and a transaction associated with items are completed.

A computing device for handling video content analysis, comprises a preprocessing module, for receiving a first plurality of frames and for determining whether to delete at least one of the first plurality of frames according to an event detection, to generate a second plurality of frames according to the determination for the first plurality of frames; a first deep learning module, for receiving the second plurality of frames and for determining whether to delete at least one of the second plurality of frames according to a plurality of features of the second plurality of frames, to generate a third plurality of frames according to the determination for the second plurality of frames; and a second deep learning module, for receiving the third plurality of frames, to generate a plurality of prediction outputs of the third plurality of frames.

A method to detect a roadway edge includes calculating a first likelihood of a roadway edge from an aerial image of a roadway by shifting a centerline of the roadway perpendicular to the centerline and overlapping the centerline with image gradients. A second likelihood of the roadway edge is determined using a vehicle telemetry fitting a probability distribution to telemetry points along the roadway. The first likelihood of the roadway edge and the second likelihood of the roadway edge are fused to identify a final likelihood of the roadway edge.

An image processing method, an image processing apparatus, an electronic device, and a computer-readable storage medium, relating to the technical field of image processing are provided. The image processing method may include performing blur classification on pixels of an image to obtain a classification mask image; and determining a blurred area of the image based on the classification mask image.

In some examples, one or more processors may receive at least one first visible light image and a first thermal image. Further, the processor(s) may generate, from the at least one first visible light image, an edge image that identifies edge regions in the at least one first visible light image. At least one of a lane marker or road edge region may be determined based at least in part on information from the edge image. In addition, one or more first regions of interest in the first thermal image may be determined based on at least one of the lane marker or the road edge region. Furthermore, a gain of a thermal sensor may be adjusted based on the one or more first regions of interest in the first thermal image.

A pole classing system may include an array of three-dimensional (3D) scanners, programmable logic controller equipment, and a computing device. The computing device may control the array of the 3D scanners to generate images of a pole from different directions and positions, generate a first pole dataset comprising dimensions and features of the pole, determine a class for the pole based on the dimensions of the pole and a first set of pole standard parameters, generate a second pole dataset by selecting a partial first pole dataset, and transmit the second pole dataset to the programmable logic controller equipment. The programmable logic controller equipment may process, based on a second set of pole standard parameters, the second pole dataset, the images and the features of the pole to thereby optimize the determined class of the pole and generate an updated class of the pole.

An autonomous vehicle is described herein. The autonomous vehicle generates segmentation scenes based upon lidar data generated by a lidar sensor system of the autonomous vehicle. The lidar data includes points indicative of positions of objects in a driving environment of the autonomous vehicle. The segmentation scenes comprise regions that are indicative of the objects in the driving environment. The autonomous vehicle generates scores for each segmentation scene based upon characteristics of each segmentation scenes and selects a segmentation scene based upon the scores. The autonomous vehicle then operates based upon the segmentation scene.