381 results about "Image selection" patented technology

The invention relates to the technical field of lung perfusion, in particular to an AI-based lung perfusion evaluation system, which comprises an image quality screening module, a blood vessel positioning analysis module, a blood flow velocity measurement module, a blood vessel anomaly analysis module and an anomaly response module, and is characterized in that the image quality screening module monitors the pixel density of an image and the variation amplitude of color gradient. According to the method, through real-time image definition evaluation and screening, the image selection process is optimized, it is ensured that all the images used for analysis reach the high-quality standard, the gray level change and the edge contour of the blood vessel are automatically calculated, the potential lesion area is accurately positioned and marked, more detailed blood vessel structure analysis is provided, and the accuracy of blood vessel analysis is improved. In addition, the system can rapidly analyze the speed deviation of the blood vessel segment, timely identify the abnormal blood flow, effectively improve the judgment speed of diseases such as pulmonary embolism or pulmonary hypertension, enhance the efficiency of coping with emergency medical conditions through automatic abnormal detection and recording, improve the accuracy and speed of judgment, and support more effective clinical decisions.

The invention discloses a visual large model Token adaptive optimization method, system and device based on differential evolution and a medium, and the method comprises the steps: carrying out the data processing of an image classification data set, an instance segmentation data set and a saliency target detection data set, and obtaining all Tokens corresponding to each image through a Patch Embedding and position coding method; obtaining a plurality of groups of Tokens corresponding to each image through a random selection mode, and performing data processing to output all Tokens corresponding to each image and the plurality of groups of Tokens selected from each image; constructing a Token adaptive selection module, a self-attention optimization module and a downstream task output module; a complete Token adaptive optimization visual large model is constructed; training a reconstruction model and a complete Token self-adaptive optimized visual large model; performing model reasoning to obtain an image classification result, an instance segmentation result image and a saliency target detection result image; the system, the equipment and the medium are used for implementing the method. The method can be widely applied to various visual tasks such as image classification, instance segmentation and saliency target detection.

The invention discloses a to-be-detected object surface image acquisition device, an image selection method and a defect detection method. The to-be-detected object surface image acquisition device comprises a projector, a diffusion plate, a reflector and a camera. The projector is suitable for projecting stripes; the diffusion plate is suitable for diffusing fringes projected by the projector to the surface of the to-be-measured object; the reflector moves between a first position and a second position relative to the to-be-measured object, and at the first position, the reflector shields a light path from the projector to the diffusion plate. And at the second position, the reflecting mirror avoids a light path from the projector to the diffusion plate. The camera is suitable for receiving a signal reflected by the surface of the to-be-measured object after the diffusion plate enters the surface of the to-be-measured object so as to obtain a phase deflection technology (PMD) image, or receiving a signal diffused by the surface of the to-be-measured object after the reflection mirror enters the surface of the to-be-measured object so as to obtain a fringe projection profile measurement (FPP) image. A phase deflection technology and a fringe projection profile measurement technology are fused in one set of defect detection system, rapid switching of two three-dimensional measurement defect detection technologies is realized for products with complex materials and different surface roughness, and the defect detection system is ensured to be high in efficiency, high in detection capability, wide in applicability and high in detection rate.

The invention discloses a screen orange peel detection method and system based on a phase deflection technology. The method comprises the following steps that a camera collects images in a stripe light source image selection and projection mode; distortion correction is carried out on the image acquired by the camera; performing phase extraction and unwrapping on the corrected image; according to the unwrapped phase data, performing three-dimensional reconstruction and gradient calculation on the image to obtain image surface height and gradient information thereof; curvature calculation and defect feature extraction are carried out according to the surface height distribution and gradient information of the reconstructed image; according to the curvature and defect feature information, the orange peel defects are quantitatively evaluated, rapid, high-precision and non-contact orange peel defect detection is achieved, the detection efficiency and accuracy are improved, and the detection cost is reduced.

The invention relates to the field of machine learning and image processing, provides an image restoration data set distillation method based on uniform image selection and neural representation, and solves the problems of large calculation amount, poor distillation performance and the like of an existing matching-based method. The method comprises the steps of training a restoration model in an original data set; based on the restoration model, calculating peak signal-to-noise ratios of restoration results of the low-quality samples in the training set and sorting the peak signal-to-noise ratios; removing partial images with the lowest peak signal-to-noise ratio and the highest peak signal-to-noise ratio, and uniformly selecting the images according to the compression ratio; respectively training the corresponding neural representation of each selected image; the mask nerve represents a part of pixels with the minimum reconstruction loss; reconstructing a recovery data set based on the mask coordinates, the distillation data and the neural representation; and training a deep learning model on the reconstructed data and testing the restoration performance. The method has the advantages that the image restoration model is efficiently trained, the storage space of an image restoration data set is greatly reduced, and iterative updating of a restoration algorithm is accelerated.

Computing an output image of a dynamic scene. A value of E is selected which is a parameter describing desired dynamic content of the scene in the output image. Using selected intrinsic camera parameters and a selected viewpoint, for individual pixels of the output image to be generated, the method computes a ray that goes from a virtual camera through the pixel into the dynamic scene. For individual ones of the rays, sample at least one point along the ray. For individual ones of the sampled points, a viewing direction being a direction of the corresponding ray, and E, query a machine learning model to produce colour and opacity values at the sampled point with the dynamic content of the scene as specified by E. For individual ones of the rays, apply a volume rendering method to the colour and opacity values computed along that ray, to produce a pixel value of the output image.

A multi-selection shutter camera application and method selectively sends images to different artificial intelligence and innovative platforms for fast sharing and informational purposes. An electronic device with a touch sensitive display and a processor is utilized. A capture screen on the display includes a multi-shutter view (a live view and at least two selective capture buttons (e.g., shutters)) for capturing media content. The method may include receiving input from the buttons to capture the content and direct it to an artificial intelligence platform; processing the content information based on the selected button; analyzing the content through parameters and show options to the user on the same screen that displays the content; and presenting a plurality of selectable options related to the user's selected shutter and intent of capturing the content, including, but not limited to uses related to at least one of the following: discovery, shopping and sharing functionality alternatives.

The invention relates to an image phase shift evaluation and phase calculation method and system based on an image selection mechanism, and belongs to the technical field of optical measurement. The method comprises the following steps: acquiring an original video, and obtaining a reference interferogram group through an image acquisition module based on the original video; obtaining an initial position interferogram based on the reference interferogram group, optimizing the recombined interferogram based on the initial position interferogram through an anti-vibration algorithm, and outputting a recombined interferogram anti-vibration sequence; calculating a phase shift accuracy index value based on the recombined interferogram sequence, presetting a phase shift accuracy index threshold, and extracting and outputting the recombined interferogram sequence of which the phase shift accuracy index value is greater than the phase shift accuracy index threshold as the target interferogram group; and calculating a phase shift quantity of the target interferogram group through a phase calculation module, calculating a phase shift evaluation index based on the phase shift quantity, and judging and outputting final phase information through the index. And image phase shift evaluation and phase calculation based on an image selection mechanism are realized.

The invention discloses an image editing method and device and a storage medium. The method comprises the steps that diffusion processing is conducted on an original image through a diffusion model; determining the total number of noise reduction iterations by using a first model trained in advance; and in each noise reduction processing process, sampling a network module of the noise prediction network in the diffusion model by using a pre-trained second model, and carrying out noise reduction processing by using the network module selected by sampling to obtain an edited target image. By applying the scheme of the embodiment of the invention, the first model determines the total number of iterations of noise reduction, and the second model samples the network module of the noise prediction network, so that the proper number of iterations and the proper network module can be adaptively selected for different images, and the noise reduction efficiency is improved on the basis of ensuring the image quality. The complexity of a long iteration process and a network structure is greatly avoided, and the generation efficiency of image editing is effectively improved.

In some embodiments, a method can include executing a first machine learning model to detect at least one lane in each image from a first set of images. The method can further include determining an estimate location of a vehicle for each image, based on localization data captured using at least one localization sensor disposed at the vehicle. The method can further include selecting lane geometry data for each image, from a map and based on the estimate location of the vehicle. The method can further include executing a localization model to generate a set of offset values for the first set of images based on the lane geometry data and the at least one lane in each image. The method can further include selecting a second set of images from the first set of images based on the set of offset values and a previously-determined offset threshold.

The invention discloses a dynamic holographic VCSEL chip based on current addressing mode multiplexing, a design method, a dynamic holographic display system and application. According to the current addressing mode multiplexing-based dynamic holographic VCSEL chip implementation method, a multi-mode VCSEL is combined with a current addressing mode multiplexing optical digital hologram, so that an ultra-compact holographic system and an ultra-high switching frame rate are realized. A holographic image is coded using an orthogonal OAM state, and a dominant OAM mode component with a maximum weight coefficient is selected to minimize channel crosstalk. The hologram is directly integrated on the surface of the VCSEL through a 3D laser nanometer printing technology to form a chip comprising a hologram array. And dynamic holographic display is realized by changing the injection current. The method fully considers the crosstalk between addressable current mode multiplexing, remarkably reduces the crosstalk in the optical multiplexing holographic technology through a systematic mode selection process, and improves the quality and efficiency of holographic display and information processing.

The invention discloses a radar anti-interference method based on frequency domain modeling and image collaboration, and the method comprises the steps: S10, carrying out the preprocessing of an original echo, and finally packaging a preprocessing result into a three-order tensor form; s20, performing FNO interference feature extraction and identification, and outputting interference type probability distribution and an interference-to-signal ratio estimation value through a de-wharf; s30, selectively processing the image, and reconstructing the histogram into an SAR image in an equalization manner; s40, evaluating the credibility of the image, and outputting an imaging credibility score between 0 and 1 through a regression head; s50, generating a credibility-interference combined criterion for judging the reliability of the radar guidance state; s60, a loss function is designed, back propagation is carried out through a total loss function, the model gradient is updated, and model parameters are optimized; and S70, dynamic anti-interference strategy matching: retrieving candidate anti-interference actions from a preset strategy library based on the interference types, performing grading and sorting according to complexity, and when the joint risk score is lower than a set threshold value, activating the optimal high-complexity anti-interference process of the corresponding interference type.

The invention discloses an AOI back inspection reflection inhibition method based on black flannelette pasting, and relates to the technical field of AOI detection, and the method comprises the steps: placing a to-be-detected silicon wafer on a silver gray aluminum material carrying platform; starting the AOI equipment to collect a background image; black flannelette is selected as a reflection inhibition material; black flannelette is attached to the center detection area of the silver gray aluminum material carrying table, and the light source perpendicular incidence area is of a double-layer structure formed by combining a bottom-layer reflective sticker and surface-layer black flannelette. After surface mounting is completed, the background image is collected again; and performing defect identification on the acquired background image, and outputting a detection report. According to the invention, through the physical light absorption characteristic of the black flannelette, the average gray value of the reflection area of the carrying platform is reduced, the high-intensity mirror reflection interference is effectively eliminated, the difference between the background gray level and the edge gray level of the silicon wafer after mounting is obvious, image analysis software can realize accurate edge searching by setting the minimum identification gray level threshold, and meanwhile, the edge searching precision is improved. The black flannelette can improve the background gray scale stability and reduce the interference of gray scale fluctuation on a threshold segmentation algorithm.

In an image selection method according to the present invention executed by a processor using a memory, a non-detection degree based on a count of undetected objects in each of a plurality of unlabeled images is predicted by inputting the plurality of unlabeled images to an undetected object prediction model (S30), and one or more unlabeled images, to which labels are to be imparted, are selected from among the plurality of unlabeled images on the basis of the non-detection degree of each of the plurality of unlabeled images (S40).

A system includes: a model configured to determine uncertainties of results of a task based on images, respectively; a classification module configured to selectively classify the images into a first category or a second category based on the uncertainties, respectively; an embedding module configured to: determine first embeddings based on the images using an embedding function; determine second embeddings for textual explanations of errors, respectively, using the embedding function; a clustering module configured to: cluster first ones of the first embeddings for images classified in the first category into first clusters; cluster second ones of the first embeddings for images classified in the second category into second clusters; and an explanation module configured to: determine similarities between each of the first and second clusters and each of the textual explanations; and determine k of the textual explanations for one of the second clusters based on the similarities.

A system generates item images using an item image generation model. The system receives a prompt for the model. The prompt is configured to request the model generate item images for an item. The system executes the model using the prompt to generate a set of item images. The system evaluates each of the set of item images to determine performance data of each of the set of item images. The system iteratively improves the set of item images by performing the following steps. The system updates the prompt based on the performance data of each of the set of item images to obtain a new prompt. The system executes, using the new prompt, the model to generate a new set of item images, and the system evaluates the new set of item images to determine performance data of each of the new set of item images.

The invention relates to a citrus surface defect detection method based on a Retinex algorithm and deep learning, and the method comprises the following steps: collecting citrus images of different defect types under different illumination conditions, and marking defect regions and defect types; adjusting the illumination uniformity and detail contrast of the citrus image by using a Retinex algorithm to form a Retinex enhanced image; introducing an attention mechanism of Transform, and performing feature extraction and attention modeling in combination with a convolutional neural network; predicting illumination distribution of the Retinex enhanced image through a deep learning model, and correcting illumination deviation; noise or artifacts in the Retinex enhanced image are removed, and an image after dynamic preprocessing is obtained; and selecting a deep learning model for defect detection, and training a defect detection model by using the dynamically preprocessed image. According to the method, the traditional advantages of image enhancement and the adaptive ability of deep learning are combined, an efficient and accurate solution is provided for citrus surface defect detection, the problems of uneven illumination, shadow interference and the like are effectively solved, and the robustness of defect detection is improved.

The invention relates to the technical field of image classification, in particular to an image classification method and system based on a visual language large model, and the method comprises the following steps: 1, obtaining a plurality of original images, and constructing an image classification network; 2, selecting one original image from a plurality of original images, inputting the selected original image into an image classification network, and finally obtaining a category prediction result; 3, constructing a loss function by utilizing a category prediction result and a real category; 4, circulating the steps 2 and 3, minimizing the loss function until the loss function converges or the number of iterations reaches a set number of times, and updating the weight of the image classification network to obtain a trained image classification network; and 5, deploying the trained image classification network to a device end, and classifying the images by using the device end to obtain a classification result. According to the method, the problems of insufficient global information capture and low visual and language information fusion efficiency in a traditional single-mode classification method are solved, and higher classification precision and task generalization ability are realized.

The invention discloses an under-forest autumn phenology quantification method and system based on an infrared camera, and belongs to the technical field of ecological detection, and the method comprises the steps: S1, under-forest vegetation image collection: employing the infrared camera to carry out the under-forest vegetation image shooting and collection at a phenology stage; s2, obtaining a daily-scale vegetation optimal image: carrying out image feature obtaining through an under-forest vegetation daily optimal image selection algorithm fusing multi-feature clustering and hierarchical screening; s3, checking the consistency of the meteorological data of the under-forest infrared camera and the data of the canopy meteorological station; s4, vegetation phenology dynamic extraction: adopting a dynamic quantile method to obtain a time sequence change curve, and determining a change rate curvature and a remote sensing phenology period of a dynamic threshold value based on a vegetation index; according to the method, forest under-forest vegetation dynamic monitoring is carried out by combining dynamic threshold identification and infrared camera imaging, redundant or low-quality images are effectively eliminated, the vegetation index extraction efficiency is improved, and the phenological period judgment precision is improved; and meanwhile, the algorithm is lightweight and efficient.

The embodiment of the invention provides an infrared and visible light image fusion method and system based on a pseudo-supervised generative adversarial network, and a medium, and the method comprises the steps: inputting an infrared image and a visible light image, carrying out the foreground and background segmentation, selecting a distance measurement mode based on the infrared segmentation image and the visible light segmentation image, and calculating the distance, performing weighting processing on the infrared distance image and the visible light distance image, and fusing the processed images to obtain a real pseudo ground image; performing analysis and fusion enhancement on the infrared image and the visible light image based on a pseudo-supervised generative adversarial network to obtain an enhanced fusion image; performing comparative analysis on the enhanced fusion image and the pseudo ground real image to obtain a judgment result, and adjusting parameters of the pseudo-supervised generative adversarial network based on the judgment result; by introducing a pseudo ground real image and a pseudo supervision generative adversarial network, the image is subjected to fusion weight fusion, and network parameters are dynamically adjusted, so that the quality of the fused image is remarkably improved.

The invention relates to an infrared small target detection method based on semantic alignment fusion and content-driven dynamic enhancement. The infrared small target detection method comprises the following steps: step 1, extracting size and edge information in an image; 2, adaptively selecting the most suitable enhancement operation based on the size and edge information extracted in the step 1, and enhancing the image; 3, on the basis of the enhanced image obtained in the step 2, different features are effectively fused through a semantic alignment fusion module to enhance image information; and step 4, using the features after semantic alignment in the step 3 to realize infrared small target detection through a decoder. According to the infrared small target detection method based on semantic alignment fusion and content-driven dynamic enhancement, a joint optimization strategy is adopted, appropriate image-level enhancement operation can be selected for each image in a self-adaptive mode, and meanwhile self-adaptive amplification is carried out on the original image. The algorithm can effectively detect the infrared small target and significantly reduce the omission ratio.

The invention discloses a fire hazard prediction method and system based on an offshore platform fire alarm device, and the method comprises the steps: determining a temperature abnormal region in a target region through a preset region division strategy, and shortening a region needing to be predicted more accurately, thereby reducing the data size needed by the subsequent fire hazard prediction, and improving the prediction efficiency. And selecting at least one target fusion image from the fusion image sequence according to a preset image selection strategy, inputting the at least one target fusion image into a preset image anomaly recognition model, outputting a recognition result corresponding to the at least one target fusion image, and performing image anomaly recognition according to each recognition result. And adopting a preset prediction strategy to generate fire prediction information of the temperature abnormal region. According to the method, the subsequent fire risk can be estimated based on the number of articles with fire hazards in a continuous period of time, so that the convenience and efficiency of fire prediction are effectively improved, and the problem that existing fire prediction is easily interfered by the environment and misinformation is generated is solved.

A system, method, and computer program product for generating a resulting image from a set of images is disclosed. The method comprises receiving an image set that includes a first image of a photographic scene based on a first set of sampling parameters and a second image of the photographic scene based on a second set of sampling parameters, and generating a resulting image based on the first image and the second image according to depth values in a selection depth map. Each distinct depth value in the selection depth map corresponds to a different image in the image set.

Described herein are systems, methods, devices, and other techniques for comprehensive and automated evaluation of digital images generated from artificial intelligence (AI) models in order to promote accurate representations of real-world content. Prompts are received at the system that are then passed to both a search engine and a generative AI model. Synthesized digital images are obtained from the generative AI model. The top-matching image from the search engine is used as a verification of the ground truth of the synthesized digital images. A realism score is generated for each synthesized digital image that characterizes the accuracy of the synthesized digital image with reference to the verification image. The realism score can be used to assist and expedite the image selection process, as well as serve as input to fine-tune the performance of generative models.

A media application receives an initial image, user input that selects one or more objects in the initial image, and a textual request to generate an output image that modifies the one or more selected objects in the initial image. The media application generates a user-selected mask that includes object pixels corresponding to the one or more selected objects. A diffusion model receives the textual request to generate the output image, a depth map, and the user-selected mask, where the diffusion model is trained to generate output pixels that are not associated with a human subject. The diffusion model outputs the output image that satisfies the textual request.

The invention provides a traditional pattern image style migration system and method based on multi-modal feature decoupling and dynamic propagation, and relates to the technical field of computer vision and digital media, and the method comprises the steps: collecting color features, texture features and vectorization composition features of a Tujia brocade image, and selecting a standard image; based on a low-rank adaptive technology, training a low-rank adaptive model by using the characteristics acquired by the Tujia brocade image; splitting an image to be migrated into migrated images according to frames, selecting key frame images, and obtaining key frame migrated images through the trained low-rank adaptive model; establishing a Tujia brocade image feature difference evaluation system, and determining a Tujia brocade image style range based on the system; and calculating the style offset condition of the key frame migration image and the standard Tujia brocade image, selecting a final fusion image from the key frame migration image according to the style range of the Tujia brocade image, and processing the final fusion image through a Poisson fusion technology to obtain a style migration image.

A system, method, and computer program product for generating a resulting image from a set of images is disclosed. The method comprises receiving an image set that includes a first image of a photographic scene based on a first set of sampling parameters and a second image of the photographic scene based on a second set of sampling parameters, and generating a resulting image based on the first image and the second image according to depth values in a selection depth map. Each distinct depth value in the selection depth map corresponds to a different image in the image set.

The invention relates to the technical field of visual inspection of photovoltaic modules, in particular to a line scanning solar module image self-adaptive splicing method based on a battery piece grid structure. The method comprises the following steps: synchronizing sub-images of the solar module and corresponding cell grid parameters, and extracting key points of each sub-image based on the cell grid parameters; determining a splicing mode of the sub-images, and calculating a position posture of each sub-image based on the key points and the splicing mode; determining the placement position of each sub-image based on the position attitude; selecting different splicing effective areas for each sub-image, and generating a standard splicing image and an extended splicing image based on the placement positions and the splicing effective areas; mapping the cell grid parameters to a standard spliced image and an extended spliced image to obtain a complete spliced image; according to the invention, the adaptive splicing of the solar module image is realized, the splicing precision of the image is improved, and the splicing precision is ensured to meet the technical requirements of photovoltaic module defect detection.

This map creation device is provided with: an acquisition unit that acquires images captured by an imaging device; a selection unit that selects an acquired image as a key frame, each time the amount of movement of a vehicle exceeds a predetermined value; a feature point detection unit that detects a plurality of feature points from the images of the key frames; a feature point selection unit that excludes, from the association of feature points, feature points, among the detected feature points, for which a distance, on the images, between consecutive key frames is smaller than a predetermined value and a luminance value is greater than a predetermined value; a feature point association unit that associates the feature points between the images of the key frames on the basis of features of the feature points that have not been excluded; and a position calculation unit that calculates three-dimensional positions of the associated feature points.

The application discloses a kind of hydraulic tunnel real-time defect monitoring method and system, the method utilizes fuzzy algorithm to calculate out the danger value of the hydraulic tunnel to be detected according to the real-time environmental factors of the hydraulic tunnel to be detected, and according to the danger value of the hydraulic tunnel to be detected obtained by calculation, for the real-time acquisition of the hydraulic tunnel image selects appropriate convolution layer and autonomous intention layer, so that the tunnel image recognition model can accurately identify the defects of the real-time image of the hydraulic tunnel to be detected, improve the accuracy of defect identification, and according to the danger value calculated according to environmental factors and defect identification result, determine the maintenance strategy of the hydraulic tunnel to be detected, improve the accuracy and effectiveness of maintenance strategy, solve the problem that the prior art only tunnel defect identification is carried out to the collected tunnel image data, while lacking the influence of internal environmental factors of hydraulic tunnel, leading to unable to accurately identify the defects of hydraulic tunnel, and then reduce the accuracy and effectiveness of maintenance strategy.