15results about How to "Improve fusion effect" patented technology

The application discloses a smart community police management system and method based on Internet of Things security, and belongs to the technical field of smart communities, and solves the problem that existing system data processing cannot utilize edge computing nodes for low-delay real-time analysis, and cannot meet timely early warning requirements, the method comprising preprocessing of multi-source heterogeneous community monitoring information, analysis and processing of the community monitoring information based on edge computing, and generation of a risk propagation path in combination with front-end analysis results and heterogeneity characteristics; in the application, analysis and processing of the community monitoring information based on edge computing avoids the delay of centralized processing, can quickly trigger early warning in the early stage of a risk event, thereby gaining golden disposal time for risk event disposal, and a risk analysis model realizes a leap from single-node anomaly detection to composite risk quantification through multi-dimensional feature fusion and dynamic space-time reasoning, and realizes multi-dimensional feature fusion and dynamic reasoning capability of the risk event.

This application relates to the field of children's playground technology and discloses a children's activity area based on a suspended walkway. The activity area includes a ground surface, a suspended walkway on the ground surface, and a movable house on the ground surface, located on one side of the suspended walkway. The movable house includes a single-story building with a second-story building on top. This children's activity area based on a suspended walkway, through the movable house, the first climbing frame inside the movable house, and the slide on one side of the suspended walkway, fully utilizes vertical space. The three-dimensional design of the activity area based on the suspended walkway allows for more diverse playground facilities within a limited ground area. The activity area features various playground facilities, providing children with more challenging and fun play experiences. The suspended walkway within the park enhances the interaction and integration with the natural environment.

This invention discloses a collaborative generative recommendation method based on multimodal semantic enhancement tagging. The method includes: constructing an entity database; extracting text semantic embeddings, image semantic embeddings, and collaborative embeddings from entities in the entity database; performing semantic enhancement collaborative alignment fusion expression processing on a semantic enhancement collaborative alignment fusion expression model to obtain quantized text embeddings and quantized image embeddings, which are then stored in the entity database; extracting text semantic embeddings and image semantic embeddings from entities of interest and inputting them into the semantic enhancement collaborative alignment fusion expression model; and searching the entity database for entities or / and combinations of entities associated with the entity of interest to generate recommendation results. This invention achieves collaborative enhancement of semantic understanding and structure awareness, enabling a unified representation of semantic richness, structure awareness, and generative guidance capabilities, providing key technical support for intelligent recommendation applications.

This invention relates to a prosthesis for implantation into bone defects. Specifically, it discloses a cranioplasty component, comprising: a bone flap body for filling a defect in the skull, the bone flap body including a disc-shaped filling body and multiple stimulation units arranged circumferentially between the outer and inner disc surfaces of the filling body, and each stimulation unit being capable of protruding from the sidewall of the filling body through radial movement; a force-applying component for driving each stimulation unit to protrude radially from the sidewall of the filling body, thereby, after the bone flap body fills the defect, the force-applying component drives each stimulation unit to protrude from the sidewall of the filling body, causing the radially outer ends of the stimulation units to press against the sidewall of the defect to stimulate the bone of the sidewall of the defect; and a positioning component for attaching and positioning the bone flap body to the skull.

The invention discloses a zero sample image restoration and enhancement method and system based on text guidance, and the method comprises the steps: obtaining an original degraded image, and constructing a positive and negative text pool; respectively generating a restoration map and a degradation map through the restoration network and the degradation network; prior loss is constructed; obtaining a reconstructed restored image and a reconstructed degraded image through the restoration model and the degradation model, and calculating reconstruction loss; converting the original degraded image, the restored image and the reconstructed restored image into image embedding vectors, converting texts in the positive and negative text pools into text embedding vectors, and screening to obtain positive and negative text embedding vectors; mapping an image embedding vector of the restored image and positive and negative text embedding vectors obtained by screening to a multi-mode embedding space to calculate and compare learning loss; and constructing a total loss function based on the prior loss, the reconstruction loss and the contrast learning loss to carry out iterative training, and outputting a restored image output by the restoration model as a final result. According to the invention, the fusion capability of the multi-modal information and the image restoration and enhancement quality can be improved.

The application provides a frequency-space joint image fusion method and device and electronic equipment, and relates to the technical field of image processing. The method comprises the following steps: inputting an optical image and a SAR image into a shallow feature extraction module in an image fusion model respectively to obtain first shallow feature information of the optical image and second shallow feature information of the SAR image; inputting the first shallow feature information and the second shallow feature information into a deep feature extraction module respectively to obtain first deep feature information of the optical image and second deep feature information of the SAR image; inputting the first deep feature information and the second deep feature information into a dual-domain feature fusion module respectively to obtain first shallow fusion feature information and second shallow fusion feature information; and inputting the first shallow fusion feature information and the second shallow fusion feature information into a fusion image reconstruction module to obtain a fusion image output by the fusion image reconstruction module, so that complementary fusion of frequency domain and spatial domain features is realized, and the quality of the fusion image is improved.

Embodiments of the present disclosure provide a method for training an image inpainting model and related equipment. The method comprises: obtaining a first facial image of a first object and a second facial image of a second object, the first facial image not having a target element, and the second facial image having the target element; performing image fusion processing on the first facial image and the second facial image to obtain a first fused image of the first object having the target element; inputting the first fused image into an initial image inpainting model to obtain a predicted image in which the target element is removed; training the initial image inpainting model according to the predicted image and the first facial image, and determining the trained initial image inpainting model as the image inpainting model. The method can improve the fusion effect of the face of the first object and the target element in the first fused image, and generate a more realistic fused image; and can automatically remove the target element in the input image while keeping other parts unchanged, thereby improving the authenticity of the image inpainting model.

This invention discloses a regional interpolation method, device, and medium, relating to the field of geographic information technology. The method includes: embedding a source region and its attributes to generate an embedding vector for the source region; embedding a target region and its known attributes to generate an initial embedding vector for the target region; inputting the initial embedding vector of the target region and the embedding vectors of the source regions within a preset spatiotemporal neighborhood of the target region into a trained regional interpolation model to obtain preliminary results for the target attributes of the target region; and processing the preliminary results according to the type of the target attributes to obtain the interpolation results for the target attributes of the target region. This effectively improves interpolation accuracy, ensures the physical rationality and numerical accuracy of the interpolation results, and enhances the applicability and robustness of the method.

This application discloses a training-free multi-image fusion method, device, and medium based on a diffusion model, relating to the field of computer vision technology. The method includes: acquiring multiple source images to be fused; simultaneously inputting the multiple source images into a diffusion model; performing a preset diffusion step on each source image to obtain a noisy image of the same size as the source image; and simultaneously performing a denoising process on the noisy image corresponding to each diffusion step until the noisy image corresponding to the 0th diffusion step is determined as the fused image of the multiple source images. This application aims to solve a series of problems in existing technologies when performing multi-image fusion, such as poor fusion effect, long processing time, and low efficiency, achieving fast and efficient multi-image fusion.

This invention provides a polyp image segmentation method based on depth guidance and a triple decoder. Using RGB images as input, the method expands the channels of the depth image, extracts its multi-layer features, and performs layer-by-layer feature fusion. This effectively utilizes the spatial structural characteristics of polyps in the depth image, enhances multi-scale semantic and spatial location information, and helps the network extract the polyp region more completely. A triple decoder architecture is designed to extract fine-grained detail features and deep semantic features of the polyps. The combined decoder integrates the advantages of both types of features, improving multi-scale feature extraction and fusion capabilities. A high- and low-resolution dual-branch feature extraction strategy is adopted to adapt to high-resolution input. Combined with the advantages of a pre-trained backbone network, this method reduces model computational complexity while ensuring no loss of polyp detail information, thus balancing segmentation performance and operational efficiency.

The application discloses a kind of based on bayesian model's unmanned aerial vehicle system oil tank area leakage accident detection method and system, through unmanned aerial vehicle carries infrared and visible light camera and shoots multimodal image of leakage scene, join geometric temperature feature and geometric priori on bayesian fusion algorithm, enhance the adaptability of this fusion model to oil tank area leakage scene, improve fusion effect and robustness;TGV regularization is used to optimize image boundary smoothness and detail retention, the model parameters are optimized by variational bayesian inference, realize the high-quality fusion of infrared and visible light image, obtain the scene situation graph of more perfect information.Further in combination with regional analysis function, the leakage area of oil tank area is automatically detected, and the leakage area features are extracted by brightness segmentation and geometric analysis, so that the leakage condition can be quickly evaluated.

The application discloses a kind of multi-modal recommendation system teacher auxiliary distillation framework and its training method, to be compressed into lightweight model with complex multi-modal recommendation model, improve system efficiency and reduce inference cost.The method comprises the following steps: the knowledge of pre-trained multi-modal teacher model is transferred to teacher helper model by first level distillation;The knowledge of the teacher helper model is transferred to student model by second level distillation to obtain lightweight recommendation model;Wherein, the parameter amount and / or computational complexity of the teacher helper model is less than the teacher model and greater than the student model;In the second level distillation, on the one hand, the difference between the teacher helper model and student model output Logits distribution is measured and minimized using Wasserstein distance, and on the other hand, a contrast learning mechanism is introduced at the feature layer to enhance the student model's ability to fuse multi-modal features.

The application relates to a sentiment analysis method, computer equipment and a storage medium, and relates to the technical field of artificial intelligence. The application can improve the accuracy of fine-grained sentiment analysis. The method comprises the following steps: acquiring multi-modal data to be analyzed, which comprises text content, non-text content and an aspect item; obtaining object-related information comprising an object text label and associated knowledge according to the non-text content; connecting the text content and the object-related information by taking the aspect item as a connection node to construct a multi-modal semantic graph; acquiring initial feature representations corresponding to the aspect item, the text content and the object-related information; inputting the multi-modal semantic graph and the initial feature representations into a graph convolutional neural network; acquiring target feature representations corresponding to the aspect item output by the graph convolutional neural network; and obtaining a sentiment analysis result of the aspect item according to the target feature representations.