Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

233 results about "Color texture" patented technology

Multi-modal entity connection method based on data enhancement

The invention discloses a multi-modal entity connection method based on data enhancement, which relates to the technical field of data processing, and comprises the following steps: enhancing entities by using a large language model, carrying out abstract processing on entity description through a design instruction, and extracting and optimizing key information; semantic information is enhanced by using a multi-modal large language model, enhanced information description is generated by fusing texts and visual features, and the enhanced information description comprises names, types, function and behavior features and physical or structural attributes of entities as well as spatial structures, appearances, color textures and other visual features extracted from the visual information; a candidate entity set is expanded and optimized through a retrieval technology, and retrieval enhancement is achieved; and performing decision selection by using a large language model, and selecting an optimal matching item from a candidate entity set obtained by retrieval enhancement. According to the method, through multi-level data enhancement and an accurate entity selection process, the accuracy and efficiency of multi-modal entity linking are remarkably improved.
Owner:CHENGDU DOCUMENT & INFORMATION CENT OF CHINESE ACAD OF SCI

Remote sensing image enhancement method and system based on visual analysis

The invention relates to the technical field of image enhancement, in particular to a remote sensing image enhancement method and system based on visual analysis. The method comprises the following steps: firstly, acquiring a plurality of preliminary connected domains of a remote sensing image through a connected domain algorithm; then, according to the color, the texture condition, the area and the shortest distance among all edge pixel points in each preliminary connected domain, obtaining a plurality of connected domains of the remote sensing image, calculating local feature indexes of the connected domains, evaluating the merging possibility, clustering and dividing earth surface feature regions, determining the pixel point window side length and the neighborhood pixel weight, and finally, obtaining the remote sensing image. And performing enhancement processing on the remote sensing image according to the weight. According to the remote sensing image enhancement method provided by the invention, the quality of the remote sensing image is effectively enhanced, the earth surface feature recognition precision is improved, the noise is effectively removed while important details are kept, the filtering result can more accurately conform to the actual features of the image, key information is kept, and the processing efficiency and the visual effect are improved.
Owner:SHANDONG GEO-SURVEYING & MAPPING INST

Forest fire prevention method based on smoke recognition

The invention belongs to the technical field of forest fire early warning, and discloses a forest fire prevention method based on smoke recognition, which comprises the following steps: acquiring heterogeneous image data from different monitoring devices, performing format unification and quality enhancement through a data preprocessing module, eliminating the difference of data from different sources, and obtaining preprocessed image data; a multi-scale feature extraction network based on deep learning is adopted, key features such as the color, texture and form of smoke are automatically learned and extracted from the preprocessed image data, and feature representation is obtained; according to the method, corresponding early warning threshold values are set according to different grades of fire judgment results, when the fire grade judgment results meet early warning conditions, early warning signals are triggered automatically, and early warning of the fire is achieved in the modes of message pushing, sound-light alarm and the like.
Owner:CHENGDU UNIVERSITY OF TECHNOLOGY

Food freshness evaluation method and system based on image processing

The invention provides a food freshness evaluation method and system based on image processing. The method comprises the following steps: acquiring a multi-view image of food, transmitting the image to a central processing center, setting personalized image acquisition parameters, and forming a food image database; based on the food image database, performing high-precision dynamic monitoring on color and texture changes of the food in different time periods, and generating a preliminary evaluation result; according to the preliminary evaluation result, combining food types, storage conditions and historical sales data, performing prediction modeling on the freshness of the food to obtain a final freshness evaluation report; and utilizing the freshness evaluation report to adjust an inventory strategy in real time, optimizing a food distribution plan, and generating an optimized management process. According to the technical scheme provided by the invention, the accuracy of food freshness evaluation is improved.
Owner:CSSC HAISHEN MEDICAL TECH CO LTD

Multi-dimensional feature and three-plane representation enhanced three-dimensional grid reconstruction method and device

The invention discloses a multi-dimensional feature and three-plane representation enhanced three-dimensional grid reconstruction method and device, and belongs to the technical field of three-dimensional reconstruction, and the method comprises the steps: obtaining three-dimensional priori knowledge: after a multi-view image sequence is generated based on a single image, extracting the multi-dimensional features of the image sequence through a multi-dimensional feature extraction and fusion module, and obtaining the three-dimensional priori knowledge; performing feature fusion on the multi-dimensional features and the initial three-plane features by using a high-resolution three-plane representation construction module containing a double-flow Transform, a convolutional layer and a pixel rearrangement operation so as to organize the initial three-plane features into high-resolution three-plane representation; a three-dimensional grid construction module is used for generating a three-dimensional grid based on the three-plane representation, a color decoder is used for decoding and generating color textures of space points based on the three-plane representation, and the three-dimensional grid and the color textures are combined to obtain a high-precision three-dimensional grid with textures. The model effectively reduces the fuzzy and artifact phenomena of the grid surface and improves the accuracy of the geometric structure. The method can be widely applied to modeling processes of game and movie and television industries.
Owner:ZHEJIANG UNIV

Pedestrian re-identification method combined with text guidance

The invention discloses a pedestrian re-identification method combined with text guidance, which comprises the following steps: collecting diversified pedestrian images, extracting image features of the images by using a deep convolutional neural network, directly generating text description from an original image through an image description model, extracting text features, and optimizing the comparison loss of the images and the text to obtain the pedestrian re-identification result. Aligning the visual features and the text features; re-recognition classification loss is optimized, it is ensured that the model can effectively distinguish identities of different pedestrians, the performance of the model is evaluated on a verification set, fine adjustment is conducted according to feedback, and the trained model is used for pedestrian recognition. According to the method, the training of the model is supervised by introducing text description, so that the limitation of a traditional pedestrian recognition technology on capturing detail features is solved. According to the method, text description (such as garment styles, colors, textures and the like) is used for guiding and optimizing the training process of a deep learning model, and it is ensured that the model can more accurately learn and recognize the detail features.
Owner:南京行者易智能交通科技有限公司

Image generation method based on style feature injection

The invention provides an image generation method based on style feature injection, which can efficiently migrate the spatial layout and color texture of a style reference image on the basis of a text image. According to the method provided by the invention, on the basis of a Stable Diffusion text map, the spatial layout and the preliminary style of an image are learned by utilizing the capability of binding features of Dreambooth, and then attention injection is executed to strengthen color features and detail textures. Therefore, the method provided by the invention has high generalization in various styles, does not need tedious parameter adjustment for a single specific style, and is highly sensitive to context features such as composition, texture, color and the like. In addition, the model needs less data, only 3-5 training images are needed, only one attention reference image is needed in the reasoning process, additional training is not needed, and the robustness is very high. Therefore, the style image generation method provided by the invention can adapt to various different styles so as to have a higher degree of freedom in a personalized task.
Owner:SICHUAN UNIV

A method for laser-induced color texture pattern on titanium surface

The invention discloses a method for laser-induced color texture pattern on titanium surface, comprising: step 1: grinding, polishing and cleaning the titanium alloy surface, drawing a pattern on the titanium alloy surface by using a laser, and continuously adjusting the parameters of the laser until a processing database in which the parameters of the laser and the color of the titanium alloy surface pattern are mutually correlated is obtained; step 2: using a colorimeter to measure and calibrate the color of the pre-drawn pattern with Lab values, calculating the color difference between the color of the pre-drawn pattern and each color in the processing database, and when the color difference is less than or equal to a predetermined value, selecting the parameters of the laser corresponding to the corresponding color for processing; the invention induces the generation of oxide film layers of different thicknesses and types on the titanium surface by a laser heat source with adjustable pulse width, makes the film layer present different colors through the interference, reflection and absorption of light, adjusts the spot size, controls the scanning speed, and is equipped with a high-speed galvanometer scanning system to realize the preparation of various fine patterns.
Owner:XIAN QINTI ZHIZAO TECH CO LTD

Railway fastener defect detection method and system

The invention relates to the technical field of defect detection, and particularly discloses a railway fastener defect detection method and system, and the method comprises the steps: controlling a high-speed industrial camera to scan a track region so as to collect a track image, carrying out the image preprocessing of the track image, extracting an ROI image block containing a potential railway fastener based on template matching, and then carrying out the detection of the potential railway fastener. The method further introduces a target detection algorithm to perform fastener accurate positioning and segmentation on the ROI image blocks to obtain finer fastener ROI image blocks, further extracts visual features of the fastener ROI image blocks, and performs feature space distribution saliency enhancement on the visual features to enhance expression effects of the visual features such as surface color, texture and shape of the fastener, so as to improve the image quality of the fastener. Therefore, intelligent identification of railway fastener defects is realized on the basis. According to the method, the limitation of a traditional detection method can be effectively overcome, the reliability of railway fastener defect detection in a complex environment is improved, and the method has relatively high universality and adaptability.
Owner:CHENGDU SEIKO HUAYAO TECH CO LTD

Multi-modal lithology identification method and system

The invention discloses a multi-modal lithology identification method and system. The method comprises the following steps: acquiring a rock sample image and a rock mesoscopic image; preprocessing the obtained images, and inputting the preprocessed rock sample images and rock mesoscopic images into the trained fusion model for real-time identification; outputting a lithology identification result by the fusion model, and naming the image according to the identification result; according to the method, the lithology of the surrounding rock is identified by establishing the fusion model based on deep learning, the fusion model can fuse the color, texture, shape and other features of the rock sample image with the crystal structure, pore distribution, particle size and other features of the rock mesoscopic image, mapping is formed by corresponding rock relations, and therefore the purpose of automatic identification is achieved; the lithology intelligent identification method based on deep learning is high in identification speed, low in cost, high in accuracy and high in automation degree, and the requirements for professional knowledge and identification experience of constructors are reduced.
Owner:SHANDONG UNIV

Real-time monitoring method and system for milk powder stirring processing

The invention provides a real-time monitoring method and system for milk powder stirring processing, and the method comprises the steps: collecting a current frame image of a milk powder material in a stirring container, and obtaining a reference frame image of the current frame image before a preset time interval; calculating a space texture feature set, a time sequence color texture feature set and a dynamic flow field feature set of the current frame image; cascading the space texture feature set, the time sequence color texture feature set and the dynamic flow field feature set to obtain a high-dimensional state vector; calculating a mahalanobis distance between the high-dimensional state vector and a target uniform state cluster core in a pre-constructed state space; when the mahalanobis distance is smaller than a first threshold value, it is judged that the current stirring state is uniform mixing; when the mahalanobis distance is greater than a second threshold value, determining an abnormal state; when the Mahalanobis distance is between the first threshold and the second threshold, it is determined that mixing is being performed.
Owner:SHAANXI YATAI DAIRY CO LTD

Multi-stage screening and quality detection method for peaches

The invention relates to the technical field of fruit and peach screening, in particular to a fruit and peach multi-stage screening and quality detection method which comprises the following steps: S1, fruit and peach variety judgment: collecting multi-angle images, extracting color, texture and contour features, and outputting variety labels and parameter sets in a classified manner; s2, growth feature mapping: extracting structure key points, matching the structure key points with a variety template, and generating a growth deviation vector; s3, area perception analysis: dividing fruit surface areas based on growth deviation, and independently identifying defects and maturity; s4, grade discrimination: fusing identification results, combining varieties and deviations, and dynamically evaluating a comprehensive quality label; and S5, screening instruction generation: converting the quality label into a control instruction, and driving the sorting device to complete output. According to the invention, integrated automatic processing of intelligent identification of fruit and peach varieties, accurate detection of regional defects and grade screening control is realized, and the accuracy and flexibility of fruit and peach screening are significantly improved.
Owner:HUNAN PROVINCIAL BOTANICAL GARDEN

Multi-model-based printed cloth pattern quality monitoring method

The invention relates to a multi-model-based printed cloth pattern quality monitoring method, which comprises the following steps of S1, carrying out image acquisition by adopting a multispectral camera to obtain multispectral image data of printed cloth, and carrying out preprocessing; s2, introducing a generative adversarial network (GAN), and enhancing the image quality by using a generator in the GAN, including denoising and detail enhancement; s3, combining Swin Transform with dynamic convolution to construct an image segmentation model, and based on the image segmentation model, performing accurate segmentation of the printed cloth, and extracting each printed area; s4, according to the segmented image data, color, texture and pattern features of printing are obtained, and fusion features are constructed; and S5, constructing a printed cloth pattern detection model, identifying abnormal patterns which do not meet quality requirements based on the fusion features, and obtaining a quality score of the printed cloth according to a detection result of the printed cloth pattern detection model. According to the invention, accurate segmentation and feature extraction of the printed cloth are realized, and comprehensive quality detection is realized by fusing various feature information.
Owner:FUJIAN YUBANG TEXTILE TECH CO LTD

Otoscope image processing system for establishing database model analysis based on deep learning

The invention provides an otoscope image processing system for establishing database model analysis based on deep learning, and relates to the technical field of image processing, the otoscope image processing system comprises an acquisition layer, a processing layer, an extraction layer, a modeling layer and an analysis layer, the processing layer adopts an AI image expansion algorithm to carry out fusion splicing and interference elimination processing on acquired images, and the extraction layer adopts a mixed feature extraction method to extract color, texture and shape features of the images processed by the processing layer and introduce MPEG-7 descriptors to express the features; according to the method, otoscope images of a patient are collected, features and feature values of the images are extracted, a focus global shared image library is constructed through deep learning modeling, the focus global shared image library is compared with the otoscope images of the patient for training, and corresponding focuses are determined, so that different ear diseases can be automatically identified and diagnosed by learning and analyzing a large amount of ear image data, and the diagnosis accuracy is improved. And more accurate and timely diagnosis information is provided for doctors.
Owner:SHENSHAN MEDICAL CENT MEMORIAL HOSPITAL OF SUN YAT-SEN UNIV

Jade defect intelligent detection method and system based on machine vision and deep learning

The invention relates to the technical field of computer vision, and discloses a jade defect intelligent detection method and system based on machine vision and deep learning, and the method comprises the following steps: S1, based on a high-resolution industrial camera and a laser three-dimensional scanner, adopting a multi-mode synchronous collection strategy, and rotating a jade sample through a precise motion control system, a jade surface high-resolution two-dimensional color image and high-precision three-dimensional point cloud data are respectively obtained, and a jade multi-mode original data set is generated. A high-resolution two-dimensional color image and high-precision three-dimensional point cloud data are integrated through a multi-modal synchronous acquisition strategy, and multi-dimensional feature expression under unified coordinates is constructed, so that the limitation of a single data source is effectively overcome; an image registration algorithm and a feature pyramid network are combined with a point cloud network to perform multi-modal feature fusion, complementarity of color texture and geometric morphology information is enhanced, and image quality is optimized based on adaptive histogram equalization and non-local mean filtering.
Owner:SHENZHEN BAIHAI DIGITAL INTELLIGENCE TECHNOLOGY CO LTD

Single-frame real-time three-dimensional imaging method based on deep learning

The invention discloses a single-frame real-time three-dimensional imaging method based on deep learning. The method comprises the following steps: constructing a lightweight phase recovery network model based on a physical model; building a multi-vision structure light stripe projection system; and constructing a training data set, completing training of the lightweight phase recovery network model, and deploying the lightweight phase recovery network model by using TensorRT. Three wrapped phase images are obtained in real time through parallel reasoning of three monochromatic view angles; unwrapping the wrapped phase into an absolute phase by using a multi-view geometric constraint and three-dimensional phase matching algorithm; and generating a three-dimensional point cloud with color texture information through phase-height mapping and a color image. According to the method, the deep learning technology and a traditional fringe projection physical model are combined, rapid phase acquisition can be completed only through a single frame of image, motion interference is overcome, and online real-time high-precision three-dimensional imaging of a complex dynamic scene is achieved.
Owner:NANJING UNIV OF SCI & TECH

Image display method and device for virtual scene, medium and program product

The invention discloses an image display method and device for a virtual scene, a medium and a program product, and belongs to the technical field of virtual world. The method comprises the following steps: acquiring a first scene depth texture map and a first scene color texture map of a first image frame; based on the first scene depth texture map, obtaining a first spatial position of a vertex of the target triangular surface in a first clipping space; mapping the first spatial position to a second spatial position in a second clipping space based on the first camera parameter and a current second camera parameter; and generating a second scene depth texture map and a second scene color texture map of the second image frame based on the second spatial position and the first scene color texture map. According to the scheme, the display frame rate of the virtual scene can be improved, so that the display effect of the virtual scene is improved.
Owner:TENCENT TECHNOLOGY (SHENZHEN) CO LTD

Three-dimensional model construction method and device based on point cloud

The invention relates to the technical field of image data processing, and discloses a three-dimensional scene construction method and device based on point cloud, electronic equipment, a computer readable storage medium and a computer program product. And the generated three-dimensional model is easy to have a large gap and a poor display effect. The method comprises the following steps: acquiring original three-dimensional point cloud data, and segmenting and projecting the three-dimensional point cloud data to obtain at least one point cloud plane; obtaining color attribute information of each point in the point cloud plane and performing color texture fitting to obtain a color texture image corresponding to each point cloud plane; filling each point cloud plane based on the color texture image; and constructing a three-dimensional model based on the filled point cloud plane. According to the method, the display effect of the three-dimensional model constructed based on the three-dimensional point cloud data is improved, and meanwhile, the data volume needing to be stored is reduced.
Owner:VKINGTELE INTELLIGENT TECHNOLOGY (SHANGHAI) CO LTD +1

Simulation bait automatic coloring method and system based on 3D model

The invention relates to the technical field of computer graphics and deep learning, in particular to a simulation bait automatic coloring method and system based on a 3D model. The method comprises the following steps: acquiring an uncolored 3D model and a reference image, and generating standard data through analysis verification, curvature grid division and image compliance detection; performing color conversion, texture enhancement and multi-scale downsampling on the compliant image to construct an image pyramid; performing multi-level feature extraction and adversarial training optimization based on a pre-trained convolutional neural network and a generative adversarial network, and generating an enhanced color texture map; performing UV expansion, color mapping and normal mapping fusion in combination with the model topology, and constructing an intermediate model with physical rendering attributes; and batch color consistency verification is realized through color histogram comparison, adaptive threshold segmentation and iteration parameter adjustment, and a standard model group is generated. According to the invention, efficient and highly realistic automatic coloring of the simulated bait is realized, and color consistency and rendering quality in batch production are guaranteed.
Owner:XINJIANG JIARUI XIUYI OUTDOOR PRODUCTS CO LTD

3D target detection tracking method based on visual image and radar tensor sparse proposal fusion

The invention provides a 3D target detection tracking method based on visual image and radar tensor sparse proposal fusion, which comprises the following steps: firstly, carrying out generalization extraction on color texture information of a visual image, and establishing multi-scale semantic high-dimensional features; secondly, extracting multi-scale space high-dimensional features of a radar tensor by using SCAN, respectively mapping a learnable sensing probe to radar and visual high-dimensional feature spaces, and performing generalization sparseness on different modal features by means of multi-head deformable attention to form radar and visual proposal features; sparse proposal fusion of radar and visual proposal features is carried out, and 3D target detection is completed; and finally, carrying out mixed multi-feature cascade matching and batch track management on a detection result, and feeding back generated track time sequence information to a front-end learnable sensing probe to realize an active target detection and tracking integrated circulating progressive network based on time sequence information guidance. According to the scheme of the invention, an integrated active sensing framework of single-frame target detection and time sequence target tracking is established, and the reliability of environment target sensing by a multi-source sensor in automatic driving is enhanced.
Owner:CHENGDU CHENGYI FUTURE TECHNOLOGY CO LTD

Ancient building heritage colored drawing layer AR visual analysis method and system

The invention provides a historic building heritage colored drawing layer AR visual analysis method and system. The method comprises the steps that correlation analysis is conducted through a multi-dimensional mapping method according to feature vectors of colored drawing colors, texture details and historical information obtained through separation, and a fusion relation matrix between attributes is obtained; according to the fusion relation matrix, a virtual reconstruction algorithm is utilized to carry out spatial recombination on colored drawing colors, texture details and historical information, and a three-dimensional virtual colored drawing model is generated; for the generated three-dimensional virtual colored drawing model, performing change simulation on color and texture details in a time dimension through a dynamic rendering technology to obtain a dynamic presentation effect; according to the dynamic presentation effect, a user control interface is embedded in the interaction environment, a switching function is set for different historical information nodes, and an interactive virtual display framework is determined; according to the corrected virtual model, dynamic rendering and interaction environment configuration are executed again, and a final visual visualization result is obtained.
Owner:GUANGZHOU CITY POLYTECHNIC +1

Solder paste state detection method, device and equipment and readable storage medium

The invention discloses a solder paste state detection method, device and equipment and a readable storage medium. The method comprises the steps that a to-be-detected solder paste area image and a corresponding template solder paste area image are acquired; the to-be-detected solder paste area image and the template solder paste area image are pseudo depth images; the to-be-detected solder paste area image and the template solder paste area image are converted into a Lab color space, and the similarity between the first color texture feature of the converted to-be-detected solder paste area image and the second color texture feature of the converted template solder paste area image is determined; and determining the state of the to-be-detected solder paste corresponding to the to-be-detected solder paste area image according to the similarity. According to the scheme, the solder paste state is determined based on the color change in the pseudo depth image, the height information participates in solder paste state detection on the basis of the two-dimensional image, the accuracy of solder paste state detection is improved, dependence on a depth camera is not needed, and therefore the solder paste state detection cost can be reduced.
Owner:CHINA GRIDCOM +2

Fruit and vegetable maturity inspection and classification method based on fruit and vegetable identification

The invention discloses a fruit and vegetable maturity inspection and classification method based on fruit and vegetable identification, and the method comprises the steps: collecting a fruit and vegetable RGB image and a near-infrared image through a high-definition camera and a near-infrared sensor, carrying out the preprocessing of denoising, enhancement, segmentation and the like, and extracting a fruit and vegetable region; extracting features such as colors, textures and shapes by using a pre-trained convolutional neural network, inputting the features into a support vector machine classification model, judging the maturity of the fruits and vegetables, and classifying the fruits and vegetables into immature fruits, mature fruits and over-mature fruits and vegetables. According to the invention, the method has a self-learning capability, can optimize the classification model through quality inspection feedback, improves the classification accuracy, and provides effective technical support for the quality control and management of fruits and vegetables.
Owner:UNIV OF SHANGHAI FOR SCI & TECH

Transformer substation external damage prevention monitoring method based on multi-modal fusion

The invention relates to the technical field of power equipment monitoring, and provides a substation external damage prevention monitoring method based on multi-modal fusion, which fully utilizes the space (depth) information of laser point cloud and the color texture information of images to carry out construction machinery detection, and adopts the D-S evidence theory to carry out confidence fusion on the two detection results, so as to obtain the substation external damage prevention monitoring method. The problem of low confidence caused by insufficient light, too small target and the like when the construction machinery is detected by independently using image data is solved, and the detection precision of the construction machinery is effectively improved; by matching the laser point cloud three-dimensional coordinate information with the two-dimensional construction machinery detection result, high-precision construction machinery positioning is realized.
Owner:YUNNAN ELECTRIC POWER TESTING & RES INST (GRP) CO LTD

Image Display Method for Virtual Scene, Device, Medium, and Program Product

Image rendering techniques for use with virtual worlds and interactive media are described herein. Techniques may include: acquiring a first scene depth texture map and a first scene color texture map of a first image frame; acquiring, based on the first scene depth texture map, a first spatial position of a vertex of a target triangle face in a first clipping space; mapping, based on a first camera parameter and a second camera parameter, the first spatial position to a second spatial position in a second clipping space; generating a second scene depth texture map and a second scene color texture map based on the second spatial position and the first scene color texture map; and displaying a second image frame based on the second scene depth texture map and the second scene color texture map. Image display frame rates for a virtual scene may be improved, thereby improving visual effects.
Owner:TENCENT TECHNOLOGY (SHENZHEN) CO LTD

Multi-modal three-dimensional point cloud semantic segmentation method for noise self-adaptive filtering

The invention discloses a multi-modal three-dimensional point cloud semantic segmentation method based on adaptive noise filtering, and belongs to the technical field of automatic driving environment perception. The method comprises the following steps: firstly, acquiring data by using a laser radar and a monocular camera, and constructing a multi-modal panoramic feature tensor containing a geometric structure and color textures through projection and mapping; extracting shallow geometric distribution features through a residual context module, and extracting multi-scale environment features through an expanded residual encoder; performing global context aggregation by using a self-attention mechanism of a Transform architecture, and establishing a full-image pixel dependency relationship to make up for a convolution locality defect; in the decoding stage, a channel cross fusion attention module (CCA) is adopted to process deep semantic features and shallow jump connection features, and a channel weight mask is dynamically generated to adaptively screen effective features and suppress high-frequency noise; and finally, outputting a two-dimensional semantic segmentation result and back-projecting the result to a three-dimensional space to obtain a semantic point cloud. According to the method, through global semantic integration and local detail screening, the problems of terrain misjudgment and noise interference of severe weather (such as rain, snow and dust) in a cross-country scene are effectively solved, and the robustness and precision of automatic driving perception are remarkably improved.
Owner:BEIHANG UNIV

Visual and auditory cross-modal object material retrieval method and system

The invention provides a visual and auditory cross-modal object material retrieval method and system, and relates to the technical field of artificial intelligence. The visual and auditory cross-modal object material retrieval system comprises a data acquisition unit for acquiring visual images containing objects made of various materials and corresponding audio data; feature extraction: feature extraction is performed on the visual and auditory data, visual features include but are not limited to colors, textures and shapes, and auditory features include but are not limited to spectrum features and time domain features; feature fusion: designing a fusion algorithm, and effectively fusing visual and auditory features to form cross-modal feature representation; and retrieving a model, wherein the model can learn a mapping relationship between the cross-modal features and the material of the object. By combining visual and auditory information, the system can more comprehensively understand the characteristics of the object, the visual information can provide the appearance characteristics of the object, the auditory information can provide the dynamic characteristics about the material of the object, and the fusion is beneficial to improving the accuracy of material identification.
Owner:ANHUI VOCATIONAL COLLEGE OF ELECTRONICS & INFORMATION TECH

Colored textile fabric image retrieval method based on dynamic feature contribution degree

The invention provides a dynamic feature contribution degree-based color textile fabric image retrieval method. Innovation and improvement are carried out aiming at single feature limitation and fixed weight fusion defects existing in a traditional retrieval technology. Color, texture and shape features are separated through a multi-modal feature decoupling technology, and coupling interference between the features is eliminated; a dynamic contribution degree quantification mechanism is innovatively constructed, gradient back propagation is used for calculating the contribution proportion of features to retrieval results in real time, and weight self-adaptive adjustment is achieved. According to the method, the problem of retrieval deviation caused by complex characteristics such as color gradient and blending structure difference of the colored spun yarn fabric is solved in a breakthrough mode, weight distribution can be intelligently matched according to inquired image characteristics (for example, the color gradient fabric focuses on color characteristics, and the jacquard fabric focuses on shape edges), and the retrieval result better fits human visual perception. According to the technical scheme, the image retrieval precision of complex fabrics in the textile industry is effectively improved, and an intelligent solution is provided for production process data association.
Owner:WUHAN TEXTILE UNIV

Simulation scene image generation method, electronic device and storage medium

A simulation scene image generation method, an electronic device and a storage medium are provided. The method includes: acquiring semantic segmentation information and instance segmentation information of a white blank 3D environment model; receiving instance text information of the white blank 3D environment model, the instance text information being editable information and used for describing an instance attribute; and generating a simulation scene image based on the semantic segmentation information, the instance segmentation information, the instance text information and a pre-trained generative adversarial network. In the method, only the establishment of the white blank 3D environment model is required, so that the simulation scene image can be generated based on the semantic segmentation information and the instance segmentation information of the white blank 3D environment model, and attributes such as color, texture and illumination do not need to be refined during establishment of the scene, thereby improving a generation efficiency.
Owner:UISEE SHANGHAI AUTOMOTIVE TECH LTD

Image-text retrieval method and system based on big data

The invention discloses an image-text retrieval method and system based on big data, relates to the technical field of data processing, and solves the technical problems that the capability of dynamically adjusting screening conditions according to data characteristics and user requirements is lacked, the internal relation between images and texts cannot be fully mined, and the user requirements are difficult to accurately match. According to the method, multi-dimensional features such as colors, textures and shapes are extracted from image information, features are extracted from text information through a bag-of-word model, deep mining of image-text information is achieved, essential features of images and texts can be represented more comprehensively, the retrieval accuracy is improved, single matching, correlation calculation and screening and secondary matching are carried out, and the retrieval efficiency is improved. The text feature and image feature matching results are further combined and screened, finally, the combined retrieval results are matched and screened again, irrelevant information can be effectively eliminated through a multi-level screening mechanism, images and texts meeting requirements are accurately positioned from mass data, and the quality of the retrieval results is greatly improved.
Owner:BEIJING AUGUST MELON TECHNOLOGY CO LTD