10results about How to "Simple network structure" patented technology

The invention discloses a target detection method, system and device based on virtual and real data and a medium, and relates to the field of target detection, and the method comprises the steps: carrying out the target detection of a sea scene image through a target detection model, and obtaining the prediction types of a plurality of sea surface targets and a plurality of candidate boxes; the target detection model is obtained by supervising and training a student model through a feature map obtained by a teacher model, the teacher model is a trained neural network model, and the student model is obtained by performing knowledge distillation on the feature map of the teacher model through a multi-valued mask; and adopting a non-maximum suppression algorithm to screen the plurality of candidate frames according to the prediction categories to obtain a prediction frame of the sea surface target, and taking the prediction category of the sea surface target corresponding to the prediction frame as the category of the sea surface target. According to the invention, the speed and accuracy of target detection in a complex marine environment can be improved.

The invention provides a TTS system, a speech synthesis method and device, electronic equipment and a storage medium. The speech synthesis method comprises the following steps: acquiring a phoneme sequence corresponding to a target text; the condition features of the reference speech are extracted, the reference speech is processed into discretized coded representation, and the condition features comprise acoustic features and / or semantic features; inputting the phoneme sequence, the condition feature and the coded representation into a pre-trained speech synthesis model, and obtaining target speech which is output by the speech synthesis model and matched with the target text; wherein the speech synthesis model comprises a neural coding and decoding language model and an audio decoder connected with the neural coding and decoding language model, and the neural coding and decoding language model comprises an autoregression module and does not comprise a non-autoregression module.

The invention relates to a cross-Kubernetes cluster service access method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: respectively setting an Ingress Nginx controller in a service providing cluster and a service consumption cluster; using a NodePort mode to set an access path used for providing use permission of an Ingress Nginx controller for an external cluster of the service providing cluster or the service consuming cluster; a Kube-apiserver is used to monitor a service updating state with an annotation mark in the service providing cluster; when it is determined that the new service appears in the service providing cluster, an Ingress access rule is created in the service providing cluster, and the Ingress access rule is used for determining a target service consumption cluster with the access permission; and starting an IP address resolution plug-in in the target service consumption cluster, and accessing the new service through a domain name access mode. By adopting the method, the new service limitation can be exposed to the specified cluster, and the cross-cluster service access efficiency and reliability are improved.

The invention discloses an inspection path determination method and device for electric power inspection and electronic equipment. The method comprises the steps that inspection task information of a plurality of inspection target objects to be subjected to electric power inspection is acquired, the inspection task information at least comprises position information and inspection duration information, and the position information is used for representing the positions of the inspection target objects; an inspection path is determined based on the inspection task information by using a pre-trained deep learning model, an electric power inspection path is obtained, the pre-trained deep learning model at least comprises a first network model and a second network model, the input end of the second network model is connected with the output end of the first network model, and the input end of the second network model is connected with the output end of the second network model; the parameter number of the first network model is greater than that of the second network model, and the inspection starting point and the inspection ending point in the power inspection path are at the same position. According to the invention, the technical problem of low efficiency of determining the electric power inspection path in the prior art is solved.

The invention relates to the technical field of remote sensing image road extraction, and provides a remote sensing image road extraction method and device based on knowledge distillation, and a readable storage medium, and the method comprises the steps: constructing a teacher model and a student model, employing an encoder-decoder architecture to learn road features for neural networks of the teacher model and the student model, the number of layers of the student model neural network is smaller than that of the teacher model neural network, a middle layer is adopted for knowledge distillation, knowledge distillation of the middle layer is isomorphic distillation, and a result is output based on the model after target training. The problems that a traditional method is large in parameter quantity and high in calculation complexity are solved, real-time reasoning can be completed rapidly and accurately, and the requirement for real-time road extraction is met.

The invention discloses an impulse noise suppression method based on gain control, which comprises the following steps: S1, automatically monitoring an input original signal, and monitoring the change of the original signal in real time; s2, monitoring the intensity of the output signal through a feedback loop, and comparing the measured output signal characteristic with the original signal characteristic to obtain an error signal, namely a pulse noise signal; s3, processing the error signal by using a self-adaptive algorithm, providing an automatically adjusted gain parameter for a gain control module, and dynamically adjusting the gain; and S4, carrying out real-time monitoring and feedback on the output signal, and adjusting controller parameters to carry out real-time correction on the output signal. By setting a feedback loop and a self-adaptive algorithm, pulse noise is continuously corrected, and the quality and reliability of signal transmission are guaranteed to the maximum extent.

The invention discloses a system for fully automatically analyzing chest CT body tissue components, which relates to the technical field of medical diagnosis equipment, can automatically select L1 slices meeting requirements and quantitatively analyze the muscle and fat area and density of a target layer, and specifically comprises three modules: a positioning module for positioning a target L1 slice on a chest CT image; the segmentation module is used for segmenting skeletal muscle and adipose tissue of an axial slice corresponding to L1 on the chest CT image; and the automatic measurement module is used for automatically measuring the body composition of the clinical chest CT image. In a word, the method not only improves the efficiency and precision of measuring the human body components on the chest CT, but also fills the blank of the conventional technology of automatically screening the L1 single slice by a machine.

The application discloses a hierarchical text classification method and system based on a lightweight graph convolutional network, acquires news body text data to be classified and a label with a hierarchical structure corresponding to the news body text data; the acquired data is input into a trained text classification network, and a text classification result is output; the trained text classification network adopts a first BERT encoder to perform encoding processing on the news body text data to be classified, and obtains body text feature representation; a lightweight graph convolutional network is adopted to perform feature extraction on the label with the hierarchical structure, and label text representation is obtained; an attention mechanism layer is adopted to process the body text representation and the label text representation, and a label-aware positive sample is obtained; a second BERT encoder is adopted to process the label-aware positive sample, and label-aware positive sample representation is obtained; and a classifier is adopted to classify the body text feature representation and the label-aware positive sample representation, and a text classification result is obtained.

The invention provides a method and system for determining an image compression model based on an asymmetric architecture, and the method comprises the steps: obtaining a trained basic neural network codec; fixing parameters of the analysis encoder, the super-prior encoder, the super-prior decoder and the context model, and determining a lightweight synthesis decoder; fixing parameters of the analysis encoder and the lightweight synthesis decoder, and determining a lightweight super-prior decoder, a lightweight context model and a super-prior encoder subjected to joint training for the super-prior decoder and the context model in combination with the super-prior encoder; and determining an asymmetric lightweight image compression model according to the analysis encoder, the lightweight synthesis decoder, the super-prior encoder subjected to joint training, the lightweight super-prior decoder and the lightweight context model. According to the invention, the decoder network structure is simplified, the calculation complexity of the decoding end is reduced, the decoding efficiency is improved, and the decoder can adapt to diversified low-end equipment.

The present invention relates to the technical field of road extraction from remote sensing images. The present invention provides a method, device and readable storage medium for extracting roads from remote sensing images based on knowledge distillation, which includes constructing a teacher model and a student model. The neural networks of the teacher model and the student model adopt an encoder-decoder architecture to learn road features. The number of layers of the student model neural network is less than that of the teacher model neural network. The knowledge distillation is performed in the middle layer, and the knowledge distillation in the middle layer is isomorphic distillation. The model outputs the result based on the target training. The problem of large number of parameters and high computational complexity of the traditional method is solved, and real-time reasoning can be completed quickly and accurately to meet the needs of real-time road extraction.