Abstract: In a method of creating a learning model using a controller configured to perform pruning on a neural network, the pruning includes a first pruning process in which a pruning process is performed in units of channels of convolutional layers and a second pruning process in which a pruning process is performed in units of weight parameters.

Abstract: A learning method for performing learning of a machine learning model using unlabeled data with no labels includes: inputting the unlabeled data to the machine learning model to generate pseudo labels; performing a first selection of selecting a pseudo label for learning from the generated pseudo labels based on reliability; performing a second selection of selecting, based on image sizes of objects to which the pseudo labels are given, the pseudo label for learning from pseudo labels that are discard targets that has not been selected as the pseudo label for learning in the first selection; and performing the learning using the pseudo label for learning.

Abstract: In an image recognition apparatus, a processor performs, based on an input image and using an image recognition model, a plurality of object detection processes to detect as an object detection region a region in the input image where a recognition target object is judged to be present. In the plurality of object detection processes, a plurality of mutually different image recognition models are used. The processor generates inference result data according to the degree of overlap among a plurality of object detection regions detected in the plurality of object detection processes.

Abstract: An object to reduce a possibility that a detection target appearing in a captured image is not detected. In order to achieve such an object, an image processing device includes: a first detection unit that is configured to detect a predetermined object and a predetermined part of the predetermined object from an acquired captured image; a setting unit that sets a detection range in the captured image based on a detection result of the first detection unit in a case where only the predetermined object out of the predetermined object and the predetermined part has been detected by the first detection unit; and a second detection unit that attempts to detect the predetermined part in the detection range.

Abstract: A vehicle onboard apparatus mounted in a vehicle includes an in-vehicle processing unit that performs image recognition on an input image from a camera mounted in the vehicle. The in-vehicle processing unit performs the image recognition using an image recognition model corresponding to position information of the vehicle and additional reference information different from the position information.

Abstract: An object to reduce a possibility that a detection target appearing in a captured image is not detected. In order to achieve such an object, an image processing device includes: a first detection unit that is configured to detect a predetermined object and a predetermined part of the predetermined object from an acquired captured image; a setting unit that sets a detection range in the captured image based on a detection result of the first detection unit in a case where only the predetermined object out of the predetermined object and the predetermined part has been detected by the first detection unit; and a second detection unit that attempts to detect the predetermined part in the detection range.

Abstract: A model generating apparatus includes a measurement unit and a change unit. The measurement unit measures a size of an object appearing on an image included in data set. The change unit changes, based on a distribution of the size, a layer to be connected to a detection unit of a Convolutional Neural Network (CNN) for detecting an object appearing on the image, among layers for extracting feature included in the CNN.

Abstract: An AI training method includes inputting training data to an existing trained AI model to acquire a trained-layer feature value output from a trained layer in the existing trained AI model, merging the trained-layer feature value with a training-target layer feature value output from a training-target layer in the training-target AI model to generate a merged feature value, and inputting the merged feature value to a training-target layer subsequent to the training-target layer to generate a new trained AI model.

Abstract: In an image recognition apparatus, a processor performs, based on an input image and using an image recognition model, a plurality of object detection processes to detect as an object detection region a region in the input image where a recognition target object is judged to be present. In the plurality of object detection processes, a plurality of mutually different image recognition models are used. The processor generates inference result data according to the degree of overlap among a plurality of object detection regions detected in the plurality of object detection processes.

Abstract: A signal color determination device includes a processor configured to acquire a captured image. If a part of a plurality of signal lamps of a traffic signal is detected from the captured image, the processor determines a signal color of the traffic signal.

Abstract: A vehicle onboard apparatus mounted in a vehicle includes an in-vehicle processing unit that performs image recognition on an input image from a camera mounted in the vehicle. The in-vehicle processing unit performs the image recognition using an image recognition model corresponding to position information of the vehicle and additional reference information different from the position information.

Abstract: An image processing device for generating learning data that is used for machine learning includes a processor that obtains image data. The processor specifies an unprocessable region that is a region in which a predetermined process cannot be performed or a region in which the predetermined process is not performed in an image region of the image data, and generates image data on which the predetermined process is performed in a region except the unprocessable region in the image region, as the learning data.

Abstract: A learning apparatus for training a student model with a teacher model includes a processor. The processor computes the performance difference between the teacher and student models. The processor makes at least one of a determination, based on the performance difference, of whether to use the teacher model and a determination, based on the performance difference, of whether to change the weight coefficient in calculating the loss in the student model.

Abstract: A model generating apparatus includes a measurement unit and a change unit. The measurement unit measures a size of an object appearing on an image included in data set. The change unit changes, based on a distribution of the size, a layer to be connected to a detection unit of a Convolutional Neural Network (CNN) for detecting an object appearing on the image, among layers for extracting feature included in the CNN.

Abstract: An object to reduce a possibility that a detection target appearing in a captured image is not detected. In order to achieve such an object, an image processing device includes: a first detection unit that is configured to detect a predetermined object and a predetermined part of the predetermined object from an acquired captured image; a setting unit that sets a detection range in the captured image based on a detection result of the first detection unit in a case where only the predetermined object out of the predetermined object and the predetermined part has been detected by the first detection unit; and a second detection unit that attempts to detect the predetermined part in the detection range.

Abstract: In a travel status recording apparatus, a data acquisition unit acquires first travel status data and second travel status data indicating a travel status in relation to a vehicle. An event detection unit detects a predetermined event that occurs in relation to the vehicle based on the first travel status data acquired. In response to the predetermined event being detected, a data analysis unit analyzes the second travel status data using artificial intelligence. A recording control unit records, in a storage unit, an analysis result of the second travel status data in association with event detection information in relation to the predetermined event.

Abstract: A system according to one aspect of the present disclosure is a program updating system including a plurality of control devices installed in a vehicle, and a gateway capable of performing in-vehicle communication with the plurality of control devices. The gateway includes: a communication unit that receives a plurality of update programs for the control devices from an external device; a storage unit that stores therein the received plurality of update programs; an in-vehicle communication unit that transmits the stored plurality of update programs to the corresponding control devices, respectively; and a processing unit that prioritizes storage of a first program defined below into the storage unit over storage of a second program defined below into the storage unit. First program: update programs of a plurality of control devices that require simultaneous updating Second program: an update program of a control device that does not require simultaneous updating.

Abstract: In a travel status recording apparatus, a data acquisition unit acquires first travel status data and second travel status data indicating a travel status in relation to a vehicle. An event detection unit detects a predetermined event that occurs in relation to the vehicle based on the first travel status data acquired. In response to the predetermined event being detected, a data analysis unit analyzes the second travel status data using artificial intelligence. A recording control unit records, in a storage unit, an analysis result of the second travel status data in association with event detection information in relation to the predetermined event.

Abstract: Provided is a control apparatus configured to control updating of a control program of an on-vehicle control device that controls a target device installed in a vehicle including a plurality of power supplies. The control apparatus includes: a detection unit configured to detect states of the plurality of power supplies that supply power to the target device; and a determination unit configured to determine that updating of the control program is possible, when the states of the plurality of power supplies detected by the detection unit indicate that at least two of the plurality of power supplies are able to output power required for updating the control program.

Abstract: Provided is a control apparatus configured to control update of a control program of an on vehicle control device which controls target equipment installed in a vehicle. The control apparatus includes: a memory configured to store an update program for the control program; a determination unit configured to determine priority levels of a plurality of update programs stored in the memory: and a control unit configured to download the plurality of update programs according to the determined priority levels in descending order of the priority levels. The determination unit determines the priority levels on the basis of one or more safety indices that are associated with each of the plurality of update programs stored in the memory.