5 results about "Dual tree" patented technology

This invention discloses a motor fault diagnosis method based on a large language model, belonging to the field of motor fault diagnosis. The method includes: first, collecting three-phase current data; then, using dual-tree complex wavelet transform to extract the amplitude and phase information of multi-layer complex wavelet coefficients as signal fault features, avoiding manual feature selection; next, constructing a multi-layer feature extraction and hierarchical weighted domain adversarial collaborative training framework that integrates grouping and Browsing optimization to extract multi-scale features and enhance domain invariance, thereby improving cross-condition generalization ability; finally, aligning the extracted signal embedding and text embedding across modalities through category-level comparative learning, and fine-tuning the large language model to recognize signal modes. This invention improves fault classification accuracy while enhancing model generalization and achieving textual semantic interpretation of diagnostic results.

An electronic apparatus performs a method of decoding video data. The method includes: receiving, from the bitstream, a plurality of syntax elements associated with a coding unit, the plurality of syntax elements indicating the coding tree type of the coding unit, and whether the local dual tree mode is enabled for the coding unit; and in accordance with the determination that the coding tree type of the coding unit is a single tree, and the local dual tree mode is enabled for the coding unit: disabling the palette mode for the coding unit when the coding unit has the size equal to or less than a predefined threshold. The disabling palette mode for the coding unit may include: disabling the palette mode for both the luma component and chroma component of the coding unit, or disabling the palette mode for only the chroma component of the coding unit.

An image decoding method according to the present disclosure comprises the steps of: deriving residual samples with respect to a target block on the basis of an inverse primary transform with respect to modified transform coefficients; and on the basis of the residual samples, generating a reconstructed picture, wherein a step of deriving the modified transform coefficients may include the steps of: parsing an LFNST index on the basis of the tree-type of a current block not being a dual-tree chroma and a MIP mode not being applicable to the current block; and deriving the modified transform coefficients on the basis of the LFNST index and an LFNST matrix.

This invention relates to the field of remote sensing image compression technology, specifically to a remote sensing image compression method based on dual-tree complex wavelet convolution and a frequency dictionary entropy model. The method includes: compressing remote sensing images using a remote sensing image compression model based on dual-tree complex wavelet convolution and a frequency dictionary entropy model; the remote sensing image compression model includes a dual-tree complex wavelet convolution module, a frequency dictionary entropy model module, and a residual feature extraction module; the dual-tree complex wavelet convolution module is used to perform downsampling and upsampling feature processing on the remote sensing image, effectively removing frequency domain redundancy in the latent representation; the frequency dictionary entropy model module is used to perform frequency division on the obtained latent representation of the remote sensing image to establish a probability model; the residual feature extraction module is used to extract features from the remote sensing image, effectively capturing long-range contextual information. This invention enhances the model's image compression effect by introducing dual-tree complex wavelet convolution and a frequency dictionary entropy model, significantly improving the fidelity of remote sensing images at high compression ratios.

This invention discloses a method and apparatus for classifying frequency-hopping signals. The method includes: acquiring orthogonal signals in the same direction; performing dual-tree complex wavelet transform on the orthogonal signals to obtain wavelet coefficients of the signals at multiple different resolution scales and multi-resolution time-frequency features of the frequency-hopping signals; inputting these features into a pre-trained multi-branch joint recognition model; using the channel attention mechanism of the multi-branch joint recognition model to perform weighted fusion of the time-frequency features of each branch to obtain feature vectors; classifying the feature vectors using a deep neural network to obtain the frequency-hopping signal classification result; and the training process of the multi-branch joint recognition model includes: obtaining the model prediction result; calculating the error of the model prediction result; and optimizing the multi-branch joint recognition model based on the error to obtain an optimized multi-branch joint recognition model. Through the above methods, this invention can solve the problems of high computational complexity and large data volume in traditional frequency-hopping radiation source identification methods.