12 results about "Push processing" patented technology

The invention provides a push processing method, a related device and a medium. The method comprises the following steps: acquiring a plurality of pushing basic features; inputting the plurality of push basic features into a multi-task prediction model to obtain first probabilities corresponding to the plurality of tasks; based on the first probabilities corresponding to the tasks, determining fusion parameters through a fusion parameter determination model; based on the first probabilities corresponding to the multiple tasks and the fusion parameters, determining a post-fusion probability; and pushing to-be-pushed content to the target object based on the fused probability, the fusion parameter determination model being trained based on a distribution position of a first sample probability, corresponding to the task, of each target sample in the target sample set in the first sample probability of each sample corresponding to the task, and pushing the to-be-pushed content to the target object based on the fusion parameter determination model. The first sample probability is obtained by inputting the sample pushing basic features of the target sample into the multi-task prediction model. According to the embodiment of the invention, the accuracy of push processing is improved. The embodiment of the invention can be applied to scenes such as content recommendation and data screening.

The embodiments of the present disclosure relate to a push processing method and apparatus, and a device and a medium. The method comprises: if candidate push objects associated with the current live broadcast video include a second push object, which matches a first push object, displaying push information in a live broadcast interface of the current live broadcast video, wherein the push information is used for prompting that the current live broadcast video includes the second push object associated with the first push object, and the first push object is determined according to historical viewed videos; and if the push information meets a preset trigger condition, displaying, in the live broadcast interface, a push page for the second push object.

The application provides a live streaming processing method and system of camera video data, an electronic device and a vehicle. The method comprises the following steps: obtaining a live streaming request instruction of a user; obtaining corresponding encoded data from a data queue based on the live streaming request instruction; the encoded data is obtained by performing hard coding processing on camera video data; in an Ffmpeg live streaming processing sub-thread, the encoded data is encapsulated into a structure object, and when it is determined that the encoded data is a key frame, a flag bit parameter corresponding to the structure object is determined to obtain a structure object to be pushed to a live streaming; and the structure object to be pushed to a live streaming is transmitted to a corresponding streaming media server. The method provided by the application can obtain camera video data in a high-performance and low-latency manner and push the camera video data to a streaming media server, thereby improving the system performance of the whole live streaming process and reducing the live streaming delay, so that the system operation resources are effectively saved.

The embodiments of the present disclosure relate to a pushing processing method, an apparatus, a device and a medium. The method comprises: in response to a first preset operation, displaying an object list interface and a plurality of entry labels on a live streaming interface of a target live streaming room, wherein candidate objects displayed in the object list interface are candidate objects corresponding to the target live streaming room; and, in response to a trigger operation for an entry label, switching the candidate objects displayed in the object list interface to candidate objects in a service scenario corresponding to the entry label.

The invention provides a background article management method and system of an intelligent media system, and belongs to the technical field of data management, and the method specifically comprises the steps: carrying out the push processing of article matching pictures of different article types based on a push processing strategy, and carrying out the push processing of the article matching pictures of the article types through the push selection data of the article matching pictures of the article types and the update data of background articles; according to the method and the device, the abnormal recognition method of the associated picture data of the article type is determined, and the push control processing target in the article type is determined according to the recognition data of the abnormal picture in the associated pictures and in combination with the push processing strategy of the image recommendation recognition model of the article type, so that the recognition processing efficiency of the abnormal picture is improved.

The embodiments of the present disclosure relate to a push processing method and apparatus, and a device and a medium. The method comprises: if candidate push objects associated with the current live broadcast video include a second push object, which matches a first push object, displaying push information in a live broadcast interface of the current live broadcast video, wherein the push information is used for prompting that the current live broadcast video includes the second push object associated with the first push object, and the first push object is determined according to historical viewed videos; and if the push information meets a preset trigger condition, displaying, in the live broadcast interface, a push page for the second push object.

The invention provides a push processing method, a related device and a medium. The method comprises the following steps: acquiring a plurality of pushing basic features, wherein the plurality of pushing basic features comprise a target object feature and a to-be-pushed content feature; determining a first feature vector corresponding to each target object feature and a second feature vector corresponding to each to-be-pushed content feature; for each first feature vector, combining the first feature vector with each second feature vector, and inputting each combination of the first feature vector and the second feature vector into a combined cross network to obtain a combined cross feature vector; and inputting the combined cross feature vector into a probability prediction model to obtain a first probability of pushing the to-be-pushed content to the target object, and pushing the to-be-pushed content to the target object based on the first probability. According to the embodiment of the invention, the accuracy of push processing can be improved, and the consumption of processing resources is reduced. The embodiment of the invention is applied to big data, data processing and the like.

Embodiments of the present disclosure relate to a video pushing method and apparatus, a device, and a medium. The method includes: detecting a trigger operation on a video sharing card during displaying the video sharing card, wherein the video sharing card is determined based on a video type of a shared video and an associated sharing object of the shared video, and the video sharing card includes a plurality of candidate card regions; and in response to detecting the trigger operation on the video sharing card, determining a target card region that matches a trigger position of the trigger operation among the plurality of candidate card regions, and displaying a pushing page corresponding to the target card region, wherein the pushing page includes a video page of the shared video and / or an object information page of the associated sharing object.

The invention discloses a personnel abnormal behavior monitoring and positioning alarm method based on Beidou and vision fusion, and relates to the technical field of intelligent safety protection and public safety monitoring, and the method comprises the steps: obtaining the positioning information and attitude information of a camera unit, and the video information collected by the camera unit; performing abnormal behavior recognition on the video information to obtain an abnormal behavior pixel coordinate value and abnormal behavior category information; determining an abnormal behavior geographic coordinate value according to the abnormal behavior pixel coordinate value, the positioning information and the attitude information; and generating alarm information according to the abnormal behavior geographic coordinate value and the abnormal behavior category information, and pushing the alarm information. The Beidou positioning information, the attitude parameters and the visual identification result are deeply fused, the abnormal behavior pixel coordinates in the video image are automatically converted into the accurate geographic coordinates, the alarm information is generated within the second-level time to be pushed, and the problem that the accurate geographic position cannot be automatically provided through traditional monitoring is solved.

The present application relates to detachably installed to the imaging device provided with the ground processing box, the imaging device includes device main body and drawer, the drawer is used to carry processing box;Device main body is provided with body electrode;Drawer has first drawer wall, second drawer wall and fifth drawer wall, fifth drawer wall connects first drawer wall and second drawer wall, and the fifth drawer wall is provided with relay contact pin for contacting with body electrode;Processing box includes: box body frame, chip assembly, with the electrical contact for being electrically connected with relay contact pin;Photosensitive drum is rotatably arranged in box body frame;Grounding assembly is achieved by contacting with the ground adapter arranged outside the processing box to ground the photosensitive drum;Pusher is connected with box body frame, and pusher is used to push processing box to installation position, and the technical scheme can make the grounding of photosensitive drum more stable.

The invention provides a push processing method, a related device and a medium. The push processing method comprises the following steps: acquiring a plurality of target push basic features; inputting the plurality of target pushing basic features into a probability prediction model to obtain a first probability that the to-be-pushed content belongs to long-tail content in which the target object is interested, and training the probability prediction model based on a total loss function, the total loss function is the weighted sum of the long tail task weight of each push basic feature sample in the push basic feature sample set on the long tail task loss function of the push basic feature sample, if the sample content meets the object interest condition and the long tail sample condition, determining the long tail task weight of the push basic feature sample as a first value, and if the sample content meets the object interest condition and the long tail sample condition, determining the long tail task weight of the push basic feature sample as a second value; otherwise, the weight of the long-tail task is determined to be a second value, and the first value is larger than the second value; and pushing the to-be-pushed content to the target object based on the first probability. According to the embodiment of the invention, the quantity and quality of the pushed long-tail content are improved. The embodiment of the invention is applied to big data, artificial intelligence and the like.