A neural network based search system is provided. A first digital image is analyzed by a user device. A targeted object in the first digital image is determined based, at least in part, on (i) the characteristics of the first digital image and (ii) the features of the targeted object. A vector array is generated based, at least in part, on (i) the first digital image and (ii) the targeted object. The vector array is analyzed by the user device. The targeted object is determined based, at least in part, by the vector array. A plurality of digital images is identified based, at least in part, on the similarity of the plurality of digital images and (i) the first digital image and (ii) the targeted object responsive to identifying a plurality of digital images, a query processing is generated. The query map is generated on a user device.

A method for recognizing a key time point in a video includes: obtaining at least one video segment by processing each image frame in the video by an image classification model; determining a target video segment in the at least one video segment based on a shot type; obtaining respective locations of a first object and a second object in an image frame of the target video segment by an image detection model; and based on a distance between the location of the first object and the location of the second object in the image frame satisfying a preset condition, determining a time point of the image frame as the key time point of the video.

Example methods and systems for quality-aware continuous learning for radiotherapy treatment planning are provided. One example method may comprise: obtaining an artificial intelligence (AI) engine that is trained to perform a radiotherapy treatment planning task. The method may also comprise: based on input data associated with a patient, performing the radiotherapy treatment planning task using the AI engine to generate output data associated with the patient; and obtaining modified output data that includes one or more modifications made by a treatment planner to the output data. The method may further comprise: performing quality evaluation based on (a) first quality indicator data associated with the modified output data, and/or (b) second quality indicator data associated with the treatment planner. In response to a decision to accept, a modified AI engine may be generated by re-training the AI engine based on the modified output data.

A system for detecting license plates is described. The system receives raw data comprising images of license plates. A base version of a ground truth is prepared based on the raw data, using a generic license plate detection (LPD). The system prepares input data for training a deep learning network. The deep learning network is trained with the prepared input data. A newly trained generic LPD is formed using data generated by the existing generic LPD.

Methods and systems for improving the interactivity of media content. The methods and systems are particularly applicable to the e-learning space, which features unique problems in engaging with users, maintaining that engagement, and allowing users to alter media assets to their specific needs. To address these issues, as well as improving interactivity of media assets generally, the methods and systems described herein provide for annotation and truncation of media assets. More particularly, the methods and systems described herein provide features such as annotation guidance and video condensation.

A medical scan labeling quality assurance system is operable to generate a set of function-generated labeling data by performing an inference function upon the set of medical scans. A plurality of sets of labeling data that includes the set of function-generated labeling data is transmitted to a client device associated with an expert user via a network for display to the expert user in accordance with anonymizing the corresponding ones of a set of labeling sources that includes the inference function. A set of correction data is from the client device, wherein each correction data of the set of correction data corresponds to one set of labeling data of the plurality of sets of labeling data. Each of a set of performance score data corresponding to the set of labeling sources is generated based on a corresponding one of the set of correction data.

Systems and methods for road typology scene annotation are provided. A method for road typology scene annotation includes receiving an image having a road scene. The image is received from an imaging device. The method populates, using a machine learning model, a set of attribute settings with values representing the road scene. An annotation interface is implemented and configured to adjust values of the attribute settings to correspond with the road scene. Based on the values of the attribute settings, a simulated overhead view of the respective road scene is generated.

Systems, methods and apparatus for data collection in an industrial environment are described. The system may include a multi-sensor acquisition component, the multi-sensor acquisition component including a plurality of inputs and outputs, a plurality of sensors operatively coupled to at least one of a plurality of components of an industrial process, and each communicatively coupled to at least one of the plurality of inputs of the multi-sensor acquisition component, a sensor data storage profile circuit structured to determine a data storage profile, a sensor communication circuit communicatively coupled to the plurality of outputs of the multi-sensor acquisition component, a sensor data storage implementation circuit structured to sensor data values in response to the data storage profile; and a data marketplace circuit structured to store at least a second portion of the plurality of sensor data values on a data marketplace.

Systems and method are provided for training an autonomous vehicle. In various embodiments, a method includes: storing, in a data storage device, real world data including a sequence of images of a road environment, the sequence of images generated based on a vehicle traversing the road environment; processing, in an offline simulation environment, the sequence of images with a deep reinforcement learning agent associated with a control feature of the autonomous vehicle to obtain an optimized set of control policies; and training the autonomous vehicle based on the optimized set of control polices.

An intelligent target object detection and alerting platform may be provided. The platform may receive a content stream from a content source. A target object may be designated for detection within the content stream. A target object profile associated with the designated target object may be retrieved from a database of learned target object profiles. The learned target object profiles may be associated with target objects that have been trained for detection. At least one frame associated with the content stream may be analyzed to detect the designated target object. The analysis may comprise employing a neural net, for example, to detect each target object within each frame. A parameter for communicating target object detection data may be specified. In turn, when the parameter is met, the detection data may be communicated.