85 results about "Biomedical image" patented technology

The invention discloses a rapid polarized light acoustic imaging device which comprises an excitation source, a light beam shaping module, a rapid polarization modulation module, an energy feedback module, a rapid scanning module and an acquisition processing module. And receiving photoacoustic signals corresponding to the exciting light in four different polarization states for calculation, and extracting anisotropy information of the sample. Compared with multiple times of scanning a sample or modulating the polarization state by rotating a half-wave plate, the problems of dislocation of multiple times of polarization acquisition points and false positive calculation are avoided, meanwhile, the time interval generated by continuously modulating the laser pulses is short, and energy fluctuation among the four pulses and errors caused by fluctuation are reduced. According to the invention, the limitation of the precision and speed of polarized light acoustic microscopic imaging is overcome, the microstructure characteristics in the target sample can be rapidly, accurately and highly sensitively detected, and the method has great application prospects in biomedical imaging and high-precision material detection.

The invention provides a laryngeal cancer multi-mode prognosis prediction method and a laryngeal cancer multi-mode prognosis prediction system fusing a CT (Computed Tomography) image and a ViT model. Relates to the technical field of biomedical images. The method comprises the following steps: acquiring and preprocessing multi-modal data of a laryngocarcinoma patient; carrying out lightweight compression, redundant information screening and robustness training on the ViT model to obtain an optimized ViT model; extracting depth features of the CT image data based on the optimized ViT model, and performing multi-stage fusion on the depth features and clinical and genome data to construct a prognosis prediction model; and performing risk stratification on the patient according to a prognosis prediction result predicted by the prognosis prediction model, and outputting treatment guidance suggestions based on the risk stratification. Through ViT model optimization, multi-modal data fusion and clinical adaptation design, precise prediction and personalized treatment guidance of laryngocarcinoma prognosis are realized, and the problems of insufficient image degradation processing, low model deployment efficiency and the like in existing laryngocarcinoma prognosis prediction are solved.

Presented herein are systems and methods for classifying biomedical images for executing operations. A computing system may identify a biomedical image of a slide with a biological sample obtained from a subject at risk of a condition; apply the biomedical image to a machine learning model; generate, based on applying the biomedical image to the ML model, a classification corresponding to the biomarker associated with the condition in biological sample on the slide; and execute an operation with respect to the slide for testing of the biological sample, in accordance with the classification for the biomedical image.

The invention relates to the technical field of biomedical image processing, and discloses a multi-fluorescence immunohistochemical image intelligent analysis method based on deep learning. The method comprises the following steps: acquiring image data through image acquisition equipment, and performing data integrity verification, format conversion and the like to obtain a standard image data set; performing noise removal and brightness equalization preprocessing on the image based on a preset rule, and extracting image features by using a deep learning model; and finally carrying out region segmentation and object identification according to the features. According to the method, full-process automation of multiple fluorescence immunohistochemical images from collection to analysis is realized, and the accuracy, efficiency and reliability of image analysis are effectively improved through standardized processing, adaptive preprocessing and deep learning feature extraction and analysis of image data; powerful technical support is provided for biomedical research and clinical diagnosis, and the method has important application value.

Presented herein are systems and methods for determining scores related to metastatic recurrence of cancers in subjects using multimodal machine learning (ML) architectures. A computing system may obtain, for a subject at risk of recurrence of cancer, a dataset comprising at least one of: (i) a biomedical image of a tissue sample from an organ associated with the cancer or (ii) a text report identifying a plurality of characteristics of a tumor associated with the cancer. The computing system may apply an ML architecture to at least one of the biomedical image or the text report of the dataset. The computing system may determine, based on applying the ML architecture, a score indicating a likelihood of recurrence associated with cancer in the subject. The computing system may generate a classification of the subject for administration of a therapy for the cancer, in accordance with the score.

The invention relates to the technical field of biomedical image processing and analysis, in particular to an organ-like three-dimensional image enhancement segmentation method and system. The system comprises a data acquisition module, a light field analysis module, a physical enhancement module, a topological feature extraction module and a boundary evolution module. The system constructs a light transmission attenuation map by using deep learning, carries out voxel-level reverse illumination compensation, and carries out dynamic evolution segmentation in combination with a centripetal vector and a topological rejection potential energy field; the core is that digital transparency is realized based on a Beer-Lambert law inverse process, optical transmission physical degradation is eliminated, and brightness distribution of deep and shallow cells is consistent; according to the method, depth invariance of feature extraction is realized, and the problem that a subsequent segmentation algorithm is sensitive to depth change is effectively solved.

The invention discloses a brain age prediction method, system and device and a medium, and relates to the technical field of biomedical image analys.The method comprises the steps that firstly, precise local detail features representing the cerebral cortex and brain tissue are precisely captured, then the dependency relationship between local areas of different images is obtained based on a windowed multi-head self-attention mechanism, and the brain age prediction result is obtained; a channel attention mechanism is introduced, importance weights of different feature channels are learned, and the importance weights are applied to feature fusion, so that long-distance dependency relationships and fine structure features are accurately captured; and then based on a cross attention mechanism, guiding global detail features to focus on a region with rich local features and guiding the local detail features to focus on a most relevant local region so as to mine deeper detail features, and finally dynamically adjusting the contribution proportion of the local detail features and global context detail features in a final decision. And the features are fused into final features so as to perceive local detail features and global detail features in a deeper level.

The invention relates to a biomedical image feature fusion method based on a multi-scale heterogeneous hypergraph. The method comprises the following steps: S1, constructing a cell-level hypergraph model; s2, constructing a region-level hypergraph model; s3, constructing a sample-level hypergraph model; s4, transmission and fusion of cross-level hypergraph features: through a level graph neural network HGCN and a dynamic attention mechanism, transmission of cell-level hypergraph features-region-level hypergraph features-sample-level hypergraph features is carried out, global alignment is carried out, and global hypergraph features are obtained; and S5, constructing a histomorphological classification model by using the global hypergraph features, and outputting sample feature representation. According to the method, through multi-scale heterogeneous hypergraph modeling and cross-level dynamic feature fusion, the problems of single-scale characterization limitation and heterogeneous data high-order interaction bottleneck are solved, the cross-scale characterization capability of the biological tissue microenvironment is remarkably improved, the adaptability, generalization performance and analysis precision of a classification model to complex data scenes are enhanced, and the method is suitable for being popularized and applied. And a general framework is provided for complex biomedical analysis tasks.

Presented herein are systems and methods for detecting basal-cell carcinoma (BCC) in biomedical images of skin lesions. A computing system may identify, for a first subject at risk of BCC in a first lesion on a region of an epidermis: (i) a first biomedical image of an outer layer of the region on the epidermis of the first subject acquired in accordance with dermoscopy, and (ii) a plurality of second biomedical images of at least one inner layer in the region on the epidermis of the first subject acquired in accordance with reflection confocal microscopy (RCM). The computing system may apply the first biomedical image and the plurality of second biomedical images to a model architecture. The computing system may generate, based on applying to the model architecture, a score indicating a likelihood of BCC in the first lesion.

The invention relates to a high-fidelity super-resolution reconstruction method and equipment for microscopic images, and belongs to the technical field of industrial automation and machine vision. The image super-resolution reconstruction network constructed by the method sequentially comprises a shallow feature extraction module, a plurality of high-fidelity groups connected in series, and a residual connection and reconstruction module. Wherein a high-fidelity block in the high-fidelity group integrates a frequency sensing high-frequency branch and a global sensing low-frequency branch adopting four-direction two-dimensional selective scanning state space modeling, and image details and a global structure are cooperatively enhanced in a spatial domain and a frequency domain through adaptive fusion of a cross-gating fusion module. A pixel domain and frequency domain joint loss function is adopted for training. According to the method, the reconstruction fidelity and the detail recovery capability of the microscopic image are remarkably improved while the linear calculation complexity is kept, and the method is suitable for industrial detection, biomedical imaging and other scenes.

Presented herein are systems and methods for classifying features from biomedical images. A computing system may identify a first portion corresponding to an ROI in a first biomedical image derived from a sample. The ROI of the first biomedical image may correspond to a feature of the sample. The computing system may generate a first embedding vector using the first portion of the first biomedical image. The computing system may apply the first embedding vector to a clustering model. The clustering model may have a feature space to define a plurality of conditions. The clustering model may be trained using a second embedding vectors generated from a corresponding second portions with at least one of a plurality of image transformation. The computing system may determine a condition for the feature based on applying the first embedding vector to the clustering model.

Presented herein are systems and methods for detecting labels in biomedical images. A computing system having one or more processors coupled with memory may identify, from a data source, a biomedical image having a first plurality of pixels in a first color representation. The computing system may convert the first plurality of pixels from the first color representation to a second color representation to generate a second plurality of pixels. The computing system may identify, from the second plurality of pixels, a subset of pixels having a color value satisfying a threshold value. The computing system may detect the biomedical image as having at least one label based at least on a number of pixels in the subset of pixels satisfying a threshold count. The computing system may store, in one or more data structures, an indication for the biomedical image as having the at least one label.

The application relates to the fields of biomedical image processing and intelligent control technology, in particular to an immune cell state analysis system based on an image processing technology; the system comprises feature extraction, state evaluation, decision generation and self-adaptive correction modules; the system extracts features by using a space-time graph neural network, the core of which is to calculate visual semantic entropy based on classification probability and feature response field, and to solve decision confidence weight by combining a cell motion diffusion index; accordingly, an AI regulation instruction and a conservative instruction based on a kinetic tolerance boundary are weighted and fused to generate a final instruction and to self-adaptively calibrate boundary parameters according to an observation error; by quantifying visual uncertainty and analyzing motion characteristics, the application effectively overcomes image blurring and non-biological interference, and significantly improves the recognition precision of active cells and the system robustness in a complex environment.

This invention relates to the field of biomedical image processing technology, and particularly to a cross-modal brain region image registration method based on differential homeomorphism. The method includes: registering the original T1-w image with an NMT-averaged standard template; registering the acquired cell architecture imaging and fluorescence imaging with the registered T1-w image; downsampling the fluorescence imaging and cell architecture imaging respectively; performing intensity correction; smoothing the cell architecture imaging / fluorescence imaging; re-downsampling the smoothed cell architecture imaging / fluorescence imaging; registering the T1-w image to the intensity-corrected cell architecture imaging / fluorescence imaging through affine transformation to obtain a deformation field; registering the T1-w D99 brain atlas to the cell architecture imaging / fluorescence imaging using the deformation field; and sampling the D99 brain atlas from the cell architecture imaging / fluorescence imaging to complete the registration. The advantages are: stronger robustness, higher accuracy, and no reliance on subsequent manual correction.

The invention provides a three-dimensional vascular structure recovery device and method fusing prior information, and relates to the technical field of biomedical image processing. Comprising a data acquisition module, a data extraction module and a complex blood vessel missing structure recovery module. The method comprises the following steps: firstly, acquiring SR-CT data of a blood vessel network, and then carrying out image segmentation on the blood vessel network to obtain a binary mask; decomposing the complex blood vessel into a series of single blood vessel images according to an image segmentation result of the blood vessel network; according to a series of single blood vessel images obtained through decomposition, selecting a single blood vessel image containing a missing structure through a connected domain analysis method; a spatial transformation network based on shape prior is constructed, and a missing structure of a single blood vessel is recovered; and finally, recombining the single blood vessel image, namely the blood vessel image without the missing structure and the blood vessel image with the restored missing structure to generate a complete blood vessel network, thereby realizing missing structure restoration of the complex blood vessel network.

The present application relates generally to identifying regions of interest in images, including but not limited to whole slide image region of interest identification, prioritization, de-duplication, and normalization via interpretable rules, nuclear region counting, point set registration, and histogram specification color normalization. This disclosure describes systems and methods for analyzing and extracting regions of interest from images, for example biomedical images depicting a tissue sample from biopsy or ectomy. Techniques directed to quality control estimation, granular classification, and coarse classification of regions of biomedical images are described herein. Using the described techniques, patches of images corresponding to regions of interest can be extracted and analyzed individually or in parallel to determine pixels correspond to features of interest and pixels that do not. Patches that do not include features of interest, or include disqualifying features, can be disqualified from further analysis. Relevant patches can analyzed and stored with various feature parameters.

Presented herein is a pipeline, integrating deep learning and language models to improve efficiency and accuracy in radiologist's workflow, as well as generate robust structured data abstracting patient's MRIs. A universally compatible pipeline can be integrated across major clinical platforms and shared with various institutions, marking a pivotal step towards reducing radiologist human error.

The present invention relates to a cross-modal medical image registration method based on signed distance function collaborative segmentation, comprising: dividing a cross-modal biomedical image into a floating image and a fixed image; estimating the signed distance field of the fixed image; obtaining a trained segmentation model; training a Voxel-Morph registration model; repeating steps (2) to (4) until the reconstruction model, the segmentation model, and the registration model converge; and inputting the estimated signed distance field and the calculated true signed distance field of the floating image to be registered into the converged Voxel-Morph registration model to generate a registered floating image. The present invention uses the signed distance field of the cross-modal image as a potential structural feature to effectively collaborate on segmentation and registration tasks, transforming the cross-modal image registration problem into a modality-independent single-modality registration problem, greatly improving the registration performance of cross-modal biomedical images in small sample scenarios, and improving the accuracy of cross-modal brain registration.

The invention relates to a miniature brain imaging device for two-color fluorescence and an imaging method thereof, and relates to the technical field of biomedical imaging device.The miniature brain imaging device comprises a shell and a lens end cover which are connected, and an imaging light path and an illumination light path are sequentially arranged in the shell and the lens end cover along an imaging optical axis; the imaging light path comprises a lens group, a first dichroscope, a first optical filter and a fifth lens which are arranged in sequence, the lens group is composed of a first lens, a second lens, a third lens and a fourth lens which are arranged in sequence, and the illumination light path comprises a first light source, a second light source, a second dichroscope and a sixth lens. The first light source and the second light source are combined through the second dichroscope to form a collimated light beam, and the collimated light beam is shaped through the sixth lens and then projected to the object plane through the first dichroscope. The application has the effects that different types of neural signals or cell populations can be monitored at the same time, different states or functional modules of neurons can be distinguished, and the resolution and accuracy of data can be improved.

The present invention relates to a method for processing sensitive data, particularly biomedical images, securely, automatically, and reproducibly on a cloud computing infrastructure. The invention also discloses the device for implementing this method. The invention relies in particular on cloud computing, cryptography, biomedical imaging, pseudonymization, anonymization, and advanced signal and image processing technologies. The invention also covers a use case for such a method through the secure implementation of image processing technologies (a business application) applied to biomedical images. In one embodiment, these images are obtained from magnetic resonance imaging (MRI), specifically for applying advanced processing with the business application to map the apparent transverse relaxation rate (R2*) and perform quantitative susceptibility imaging (QSM).Figure for the abbreviation: figure 1.

The invention discloses a biomedical image segmentation method, system, device and medium, and relates to the technical field of medical images, and the method comprises the steps: collecting a biomedical image, and decomposing the biomedical image into a low frequency (LF) component and a high frequency (HF) component; a low-frequency (LF) component and a high-frequency (HF) component are respectively output to two parallel branches, a cross-branch pseudo-label guiding strategy is introduced, a pseudo-label generated by the low-frequency (LF) branch guides a high-frequency (HF) branch and captures a semantic context, and a pseudo-label generated by the high-frequency branch supervises the low-frequency branch to enhance the perception of fine-grained texture; and carrying out dynamic weight adjustment on a space quadrant of the feature map, selecting a better frequency branch as a final reasoning path, abandoning another branch, and generating a corresponding segmentation prediction result. According to the method, collaborative modeling of low-frequency semantics and high-frequency structures of the images is realized, so that the segmentation accuracy and generalization ability of the model under full-supervised and semi-supervised conditions are enhanced.

The invention discloses a method for generating a biomedical image follow-up report fused with structured quantitative features, and the method comprises the steps: obtaining original biomedical images at multiple time points, carrying out the image preprocessing, and obtaining the quantitative features through the processing of a pre-trained advanced computer vision model; constructing a high-dimensional structured quantitative feature vector based on the quantitative features; encoding the preprocessed original biomedical image to obtain an image visual code, and taking the image visual code and the high-dimensional structured quantitative feature vector as input of a visual language large model to obtain an initial biomedical image follow-up report; calculating a consistency comprehensive score based on the initial biomedical image follow-up report and the high-dimensional structured quantitative feature vector; and correcting the initial medical image follow-up visit report based on the consistency comprehensive score. According to the method, the accuracy, the consistency and the overall scientific reference value of the finally generated biomedical image follow-up visit report in key quantitative indexes and variation trend description are remarkably improved.

This invention relates to the field of biomedical image processing technology, and discloses a method and system for quantitative analysis of hypothalamic immunofluorescence images. The method includes: performing spectral unmixing on multispectral fluorescence microscopy images based on a nonnegative matrix factorization algorithm to obtain a clean signal distribution map; using Gaussian Laplace filtering and watershed transform to achieve cell detection and segmentation; performing affine and B-spline registration between slice images and standard brain atlases to generate regions of interest masks for neural nuclei; using a local background adaptive correction strategy to perform fluorescence quantification and positive determination; and calculating Pearson correlation coefficient and Manders overlap coefficient to achieve colocalization analysis. The system includes a spectral unmixing module, a cell detection and segmentation module, an atlas registration and region recognition module, a fluorescence intensity quantification module, and a colocalization analysis and statistical output module.

The present invention discloses a non-interference, non-iterative complex amplitude reading method and device. The reading method comprises the following steps: diffracting a light beam containing amplitude and phase information to obtain an intensity image as a diffraction pattern with varying light intensity; constructing and training a diffraction intensity-complex amplitude model based on the correlation between the diffraction pattern and the amplitude and phase information; and directly applying the trained model to new diffraction images to obtain amplitude and phase information; and being able to detect complex amplitude information including amplitude and phase from an intensity image based on a single diffraction image. This method can improve the stability and accuracy of phase reading results, increase calculation speed, and simplify the optical system. The method is suitable for fields such as holographic storage, biomedical image processing, and microscopic imaging.

The application discloses a skeleton quality driven tubular structure segmentation closed loop optimization method and system, which is applied to the technical field of biomedical image processing and computer aided diagnosis, and the method comprises the following steps: using a trained segmentation model to infer medical volume data to obtain a segmentation mask; skeletonizing the segmentation mask to extract a skeleton graph; performing topological defect detection on the skeleton graph to obtain a structured defect report; reversely mapping three-dimensional coordinates in the report back to a voxel space, taking each defect point as a center, a preset radius as a range, and generating a defect density weight graph according to a severity score; and using the weight graph as a spatial weighting parameter of a loss function to optimize the segmentation model; and the application reversely maps the topological defects detected by skeletonization into a weight graph and integrates the weight graph into a loss function for closed loop iteration, realizes directional repair on high-occurrence areas such as fractures and false branches, and improves the conduction efficiency of segmentation improvement to skeleton quality.

The invention provides a biological tissue image segmentation method based on deep learning, and aims to improve the segmentation accuracy of a tissue structure in a biomedical image with low contrast, more noise and fuzzy boundary, and the method comprises the following steps: constructing a U-shaped network model containing an improved convolutional neural network structure, the model has enhanced feature extraction capability and detail recovery capability; preprocessing the biological tissue image, enhancing the contrast ratio of the image and reducing noise; outputting a segmentation result of the biological tissue image through the U-shaped network; and finally generating a high-precision biological tissue segmentation image according to the cross entropy loss function and the Dess coefficient optimization model. The method is suitable for various biomedical image data including but not limited to computed tomography (CT), magnetic resonance imaging (MRI) and tissue slice images, can effectively improve segmentation precision and reduce calculation overhead, and has wide clinical application potential.

The present invention relates to a method for processing sensitive data, particularly biomedical images, securely, automatically, and reproducibly on a cloud computing infrastructure. The invention also discloses the device for implementing this method. The invention relies in particular on cloud computing, cryptography, biomedical imaging, pseudonymization, anonymization, and advanced signal and image processing technologies. The invention also covers a use case for such a method through the secure implementation of image processing technologies (a business application) applied to biomedical images. In one embodiment, these images are obtained from magnetic resonance imaging (MRI), specifically for applying advanced processing with the business application to map the apparent transverse relaxation rate (R2*) and perform quantitative susceptibility imaging (QSM).Figure for the abbreviation: figure 1.

The invention provides an artificial intelligence-based system for generating personalized medical information through the integration of multi-modal data across pre-hospitalization, hospitalization, and post-hospitalization phases. The system employs encoder modules to process diverse input modalities, including audio recordings, video streams, biomedical images, text-based clinical records, and physiological signals. These encoded representations are integrated into a unified latent space using a large language model (LLM) trained on medical datasets comprising historical patient cases, anatomical knowledge, and treatment guidelines. The LLM enables cross-modal analysis to generate personalized outputs via decoder modules, which transform the latent space representation into actionable formats like text-based summaries, visualizations, audio explanations, and treatment suggestions. A key innovation is real-time intraoperative feedback via encoder-decoder modules detecting anatomical structures and deviations from standard protocols. The system also includes a federated learning module to aggregate model updates across medical centers while preserving patient privacy through deidentification protocols.

The invention relates to a chick sex identification model construction method, a chick sex identification method and a chick sex identification system. The chick sex identification method comprises the following steps: acquiring historical anus images at anus positions of multiple varieties of chicks, wherein the historical anus images are extracted frame by frame from a continuous video stream; preprocessing each historical anus image to obtain a historical image data set; constructing a basic chick sex identification model based on the historical image data set; and performing multi-task learning and transfer learning on the basic chick sex identification model to obtain a plurality of chick sex identification models corresponding to the plurality of varieties. According to the method, the common problem of feature loss in a high-noise biomedical image in a traditional method is solved, the problem of edge sawtooth caused by conventional clustering is avoided, and the chick sex identification accuracy of the chick sex identification model is remarkably improved.

The invention relates to a cell real-time detection method and device based on improved YOLOv12 and a storage medium, and is applied to the technical field of computer vision and biomedical image processing.The method includes the steps that an attention enhancement module AFE is arranged in a backbone network of an existing YOLOv12 model, the module combines channel attention with a space attention mechanism, and the attention enhancement module AFE is used for enhancing the attention of cells in the backbone network of the existing YOLOv12 model; a self-adaptive compression and spatial information aggregation strategy is adopted, so that the characterization capability of the model on key features such as tiny cells and fuzzy edges is effectively enhanced; an attention fusion module A2C2f is used in the neck network to replace an original standard C3k2 module at a target node, a fusion path of bottom-layer features and high-layer features is optimized, and collaborative optimization of feature extraction and multi-scale feature fusion is realized; according to the structure, the feature extraction capability of the model on tiny and fuzzy targets can be enhanced, and the fusion efficiency of multi-scale features is improved, so that the detection precision and the recall rate in complex scenes such as dense cells and overlapping are remarkably improved, and the omission ratio is effectively reduced.