The present invention makes it possible to cause an action recognizer capable of recognizing actions with high accuracy and with a small quantity of learning data to learn. An input unit 101 receives input of a learning video and an action label indicating an action of an object, a detection unit 102 detects a plurality of objects included in each frame image included in the learning video, a direction calculation unit 103 calculates a direction of a reference object, which is an object to be used as a reference among the plurality of detected objects, a normalization unit 104 normalizes the learning video so that a positional relationship between the reference object and another object becomes a predetermined relationship, and an optimization unit 106 optimizes parameters of an action recognizer to estimate the action of the object in the inputted video based on the action estimated by inputting the normalized learning video to the action recognizer and the action indicated by the action label.

An information generation device includes: a measurement unit configured to obtain a plurality of measurement results by performing a measurement of a vehicle size with respect to a same traveling vehicle a plurality of times; a detection unit configured to detect an accuracy of each of the plurality of measurement results; and a determination unit configured to determine a vehicle size of the traveling vehicle from the plurality of measurement results on the basis of the accuracies.

A vehicle can include a window. The window can include an interior side and an exterior side. The vehicle can include a camera operatively positioned to capture visual data of a portion of an exterior environment of the vehicle. The vehicle can include a dual-sided transparent display forming at least a portion of the window. The vehicle can include a processor operatively connected to the camera and the dual-sided transparent display. The processor can be configured to selectively cause the dual-sided transparent display to display exterior visual information on the exterior side. The processor can be configured to cause the dual-sided transparent display to display interior visual information on the interior side. The interior visual information can include the visual data of the portion of the exterior environment of the vehicle.

A drive-through treatment system for a drive-through clinic includes: a plurality of monitoring devices positioned at an inlet of a plurality of treatment lines; a plurality of treatment devices positioned on a plurality of medical shelters in each of the plurality of treatment lines; a plurality of outlet devices positioned at an outlet of the plurality of treatment lines; and a treatment line determining device for determining a treatment line into which a target vehicle to enter the drive-through clinic will enter among the plurality of treatment lines based on information received from a plurality of monitoring devices and a plurality of outlet devices. Each of the plurality of monitoring devices may check a treatment start, a treatment progress situation, and a treatment speed of vehicles that have entered the treatment line from the inlet of the corresponding treatment line to be transmitted to the treatment line determining device.

A smart parking management system and a smart parking management method are provided. The system includes a first smart pole, multiple second smart poles, and a processing device. An image capturing range of the first smart pole covers an entrance of a road section. An image capturing range of each second smart pole covers at least a parking space of the road section. The processing device is communicatively coupled to the first smart pole and the second smart poles. The processing device identifies first vehicle information of a first vehicle entering the road section according to an image stream captured by the first smart pole, obtains a movement trajectory of the first vehicle in the road section based on the first vehicle information and an image stream captured by each second smart pole, and determines where the first vehicle is parked according to the movement trajectory.

Systems and methods for detecting or predicting potential collisions between vehicles are provided. The systems and methods may receive sensor output indicative of a location, a heading, and/or a moving speed of a first vehicle and/or a second vehicle. The systems and methods may predict a collision between the vehicles at an intersection between routes based on the received sensor output. The systems and methods may change movement of the first vehicle and/or the second vehicle responsive to predicting the collision.

A system includes a processor and a memory storing instructions that, when executed by the processor cause the system to generate a machine learning model; generate an artificial neural network; analyze an image of a parking area using a spot detection machine learning model; analyze the image of the parking area using a vehicle detection machine learning model; and classify a parking space as available when an area of intersection does not exceed a predetermined value. A method includes analyzing an image of a parking area using a first machine learning model; analyzing the image of the parking area using second machine learning model; and classifying a parking space as available when an area of intersection does not exceed a predetermined value. A method includes generating a spot detection machine learning model; and generating, by analyzing a plurality of labeled images, an artificial neural network.

A system for classifying an object may include one or more processors, a sensor and a memory device. The memory device may include a data collection module, a starburst module, and an object classifying module. The modules have instructions that when executed by the one or more processors cause the one or more processors to obtain three dimensional point cloud data from the sensor, identify at least one cluster of points representing the object within the three dimensional point cloud data, identify a center point of the at least one cluster of points, project a plurality of rays from the center point to points of the at least one cluster of points to generate a shape, compare the shape to a plurality of candidate shapes, and classify the object when the shape matches at least one of the plurality of candidate shapes.

A first image can be acquired from a first sensor included in a vehicle and input to a deep neural network to determine a first bounding box for a first object. A second image can be acquired from the first sensor. Input latitudinal and longitudinal motion data from second sensors included in the vehicle corresponding to the time between inputting the first image and inputting the second image. A second bounding box can be determined by translating the first bounding box based on the latitudinal and longitudinal motion data. The second image can be cropped based on the second bounding box. The cropped second image can be input to the deep neural network to detect a second object. The first image, the first bounding box, the second image, and the second bounding box can be output.

An apparatus, method and computer program product are provided for determining a state of visibility of a road object, such as a road sign, using vehicle sensor data. For example, the apparatus determines whether one or more sensors of a first vehicle observes a road sign. If the road sign is not observed by the one or more sensors, the apparatus determines whether one or more second vehicles is obscuring the road sign. If the one or more second vehicles is not obscuring the road sign, the apparatus determines whether the road sign is obscured due to a weather condition. If the road sign is obscured due to the weather condition, the apparatus generates a signal indicating that the road sign was obscured due to the weather condition.