CN111652181B - Target Tracking Method And Device And Electronic Equipment - Google Patents
Legal status (The authorized standing is an assumption and iTagPro website is not a authorized conclusion. Current Assignee (The listed assignees could also be inaccurate. Priority date (The precedence date is an assumption and isn't a authorized conclusion. The applying discloses a goal monitoring method, a target tracking device and electronic gear, and relates to the technical area of synthetic intelligence. The strategy includes the following steps: a first sub-network in the joint tracking detection community, a primary function map extracted from the target function map, iTagPro smart tracker and a second function map extracted from the goal feature map by a second sub-community in the joint tracking detection network; fusing the second characteristic map extracted by the second sub-community to the first characteristic map to obtain a fused feature map corresponding to the first sub-network; buying first prediction information output by a primary sub-community based mostly on a fusion function map, and iTagPro website buying second prediction info output by a second sub-community; and determining the current position and the movement trail of the moving goal in the target video based on the first prediction information and the second prediction information.
The relevance among all the sub-networks that are parallel to one another might be enhanced by way of characteristic fusion, and the accuracy of the decided position and motion trail of the operation target is improved. The present utility pertains to the sector of synthetic intelligence, and in particular, to a goal monitoring methodology, iTagPro product apparatus, and electronic gadget. In recent times, iTagPro website artificial intelligence (Artificial Intelligence, iTagPro key finder AI) technology has been widely utilized in the sector of target tracking detection. In some scenarios, a deep neural network is typically employed to implement a joint trace detection (tracking and object detection) network, the place a joint hint detection network refers to a community that is used to attain target detection and target trace collectively. In the existing joint tracking detection community, the position and movement path accuracy of the predicted transferring goal is not excessive sufficient. The applying offers a goal tracking technique, a target tracking device and digital gear, which might enhance the issues.
In one aspect, an embodiment of the present application provides a target monitoring methodology, where the strategy contains: iTagPro website a first sub-community in a joint tracking detection community is used for extracting a first characteristic image from a target characteristic image, iTagPro smart tracker and a second sub-community in the joint tracking detection community is used for iTagPro key finder extracting a second characteristic image from the target characteristic picture, wherein the target characteristic picture is extracted from a video body of a target video; fusing the second function map extracted by the second sub-network to the first feature map to acquire a fused function map corresponding to the primary sub-community; buying first prediction data output by a first sub-network according to the fusion characteristic map, and acquiring second prediction data output by a second sub-network; based mostly on the first prediction info and the second prediction data, figuring out the present position and the motion trail of the transferring target within the target video.
Optionally, in the tactic offered by the embodiment of the present utility, the primary subnetwork is a classification subnetwork, and iTagPro website the second subnetwork is a regression subnetwork or a tracking subnetwork. Optionally, in the tactic offered by the embodiment of the current application, the first subnetwork is a regression subnetwork, and the second subnetwork is a classification subnetwork or a tracking subnetwork. In one other facet, an embodiment of the present application gives a target monitoring apparatus, including: the machine comprises a feature acquisition module, a characteristic fusion module, a prediction module and a monitoring module. The function acquisition module is used for iTagPro website detecting a first sub-community within the community through joint monitoring, extracting a primary characteristic map from a target function map, and extracting a second feature map from the goal function map by way of a second sub-network within the network by joint tracking, wherein the target feature map is a function map extracted from a video frame of a target video.
