13 results about "Color normalization" patented technology

The invention discloses a skin image abnormal region detection method based on a convolutional neural network, and relates to the technical field of medical image analysis, and the method comprises the following steps: S1, carrying out adaptive illumination and color normalization processing and two-dimensional fast Fourier transform; s2, multi-scale representation is fused in a cross-scale mode; s3, learning the dynamic weight of the multi-scale features and carrying out weighted summation; s4, through improving a DANet model, executing double-path processing of Fourier domain semantic modulation and morphological prior space attention, and gating bidirectional aggregation; s5, carrying out binarization and connected domain analysis, and extracting candidate focus areas; s6, extracting an instance-level feature vector, and estimating a corresponding cognitive uncertainty value; and S7, performing graph relation reasoning and multi-head decoding. According to the method, the limitations of neglect of association between lesions, single evaluation dimension and poor prediction generalization ability in a traditional method are effectively overcome, and an efficient and reliable solution is provided.

This invention relates to the field of image recognition and detection technology for foodborne pathogens, and discloses a method for identifying Listeria monocytogenes based on image recognition. The identification method includes: acquiring images of chromogenic culture medium plates and converting them to Lab and HSV color spaces; performing adaptive color normalization based on the background region of the culture medium; performing colony instance segmentation on the normalized image and extracting extended regions of interest; extracting saturation distribution sequences along radial rays and detecting halo transition patterns using first-order difference; statistically analyzing halo angle coverage and average transition amplitude, and combining halo quantization feature vectors; inputting the feature vectors into a gradient boosting decision tree classifier to output colony identification results. This invention solves the technical problems of unstable classifier discrimination boundaries caused by color shifts between different batches and the difficulty in detecting halos caused by weak lecithinase reactions.

This invention discloses a method and system for quantitatively assessing the degree of fibrosis in kidney pathological sections, belonging to the field of medical image processing technology. The method includes: performing color normalization and adaptive contrast enhancement preprocessing on Masson stained section images; classifying pixel-level tissue components using a multi-component semantic segmentation network containing a channel-space dual-path attention module; calculating the fibrosis area ratio and outputting a CI score based on the Banff grading standard; generating a fibrosis spatial distribution heatmap; and calculating the annual fibrosis progression rate through longitudinal follow-up comparison.

The present application relates to the technical field of image processing, and discloses a light staining sperm microscopic image light-weight style consistency preprocessing method and system, comprising: acquiring a light staining sperm microscopic image, performing color normalization on the light staining sperm microscopic image according to a deep staining sperm microscopic image, performing local contrast enhancement on the normalized light staining sperm microscopic image according to local contrast of a sperm head, a tail and a boundary region, performing brightness correction on the enhanced light staining sperm microscopic image according to the definition of the sperm head, the middle section and the tail structure and the background stability, and obtaining a light staining sperm microscopic image after light-weight style consistency preprocessing. The present application can solve the problems of staining difference, uneven brightness and boundary blur through image preprocessing without increasing the number of samples and additional training burden, improve the overall quality of the light staining sperm microscopic image, and improve the detection performance of the subsequent model.

The invention discloses a non-abandoned pattern rendering enhancement method based on generative image processing. The method comprises the steps of non-abandoned pattern collection, image preprocessing, enhancement generator model construction, discriminator model construction and non-abandoned pattern rendering enhancement. The method comprises the following steps: acquiring original image data through a non-missing pattern; image preprocessing methods of color normalization, random visual confusion, geometric space deformation, multi-scale image generation and data set segmentation are adopted; the enhancement generator model outputs a non-abandoned pattern rendering enhancement result, and the accuracy of an output image is ensured by simulating a cognitive process formed by layering patterns and integrating cultural implied meaning description and process physical characteristics at the same time; a four-level parallel heterogeneous model is adopted as a discriminator model, a multi-level identification system is constructed, and a generated image is evaluated from four dimensions of pixel authenticity, pattern structure rationality, image and text culture semantic consistency and process feature accuracy, so that high fidelity of an enhanced result is ensured.

The invention discloses an apple stem and calyx identification method based on binocular vision multi-mode fusion, which comprises the following steps: synchronously acquiring RGB images of a left view angle and a right view angle of an apple, calculating a disparity map according to the RGB images of the left view angle and the right view angle and generating a depth image, and carrying out space calibration and pixel-level registration on the RGB images and the depth image; performing color normalization, reflection suppression and brightness compensation processing on the registered RGB image, and performing hole filling and filtering smoothing processing on the registered depth image; performing multi-scale fusion on the processed RGB image and the depth image, and constructing multi-modal input data; and inputting the multi-modal input data into a semantic segmentation network based on an encoder-decoder architecture, and outputting a pixel-level semantic segmentation mask of the stem and the calyx based on the semantic segmentation network to obtain an identification result. According to the method, the problem of confusion of fruit stems / calyx and fruit defects is effectively avoided, so that the recognition precision is further improved.

The invention provides a tumor pathological image typing and grading method and device based on deep learning and a residual network technology, and relates to the technical field of image processing, and the method comprises the steps: collecting and preprocessing a tumor pathological image, segmenting a tissue region, carrying out the color normalization, and dividing the tumor pathological image into image blocks; constructing a deep convolutional feature extraction network based on a residual network, and extracting shallow and deep feature maps; performing up-sampling on the deep feature map and performing weighted fusion on the deep feature map and the shallow feature map, and inputting the deep feature map and the shallow feature map into parallel convolution branches to obtain typing and grading feature maps; calculating a feature deviation degree based on the hierarchical feature map to generate a space attention map, and determining a region focusing degree; enhancing the typing feature map by using the attention weight; respectively carrying out global pooling and classification on the enhanced feature map and the graded feature map, and outputting a typing and grading result; screening high-focus points according to an area focusing degree set threshold value, and generating a key area thermodynamic diagram; according to the invention, typing and grading cooperative processing is realized, and the diagnosis efficiency and accuracy are improved.

The invention discloses a road ecological landscape dynamic classification method and system based on landscape visual features, and relates to the field of artificial intelligence, and the method comprises the steps: constructing a data set containing image-level and pixel-level double-layer annotations; preprocessing the image through self-adaptive illumination-color normalization in combination with local statistics and global estimation; carrying out image block division and recombination by adopting semantic guidance; constructing an end-to-end classification model, deeply fusing visual and semantic features, modeling a spatial layout and semantic similarity relationship, and realizing landscape type discrimination; and utilizing the training model to automatically classify the new image. According to the method, the robustness to complex illumination can be effectively improved, structured image blocks rich in semantics are formed, multi-scale feature interaction and discrimination are enhanced, and feature representation consistent in semantics is learned.

The application discloses a drilling stratum AI logging method and system based on machine vision, collects multispectral drilling core image and multi-view ring scanning core surface image and depth identification information, generates a color consistent image group through light-color normalization registration; generates a lithology discrimination feature set through image partition splicing, lithology end member decomposition and grain size classification feature extraction; associates the lithology discrimination feature set with the depth identification information to establish a depth-lithology contrast index, identifies a stratum interface position through compaction compensation and cycle direction analysis, implements feature out-of-limit detection to generate a mutation trigger identification; extracts an interface transition zone feature slice according to the mutation trigger identification, constructs a depth-image analysis profile through multi-sampling point edge trend continuity evaluation and geometric distortion correction; finally, performs lithology partition boundary positioning and lithology classification comparison, and outputs a standardized image logging report, effectively improving the accuracy of lithology interface identification and the standardization degree of logging results.

The invention discloses a multi-literature mixed character recognition method and system based on deep learning. The method comprises the steps of text image collection, primary data processing, text detection model construction, text recognition model construction and multi-literature mixed character recognition. According to the invention, original image data is obtained through text image acquisition; a data primary processing method of color normalization, random visual confusion, geometric space deformation, multi-scale image generation and data set segmentation is adopted; a multi-granularity deep learning model is adopted as a text detection model, features of different granularities are captured through an adaptive receptive field mechanism, and sub-tasks are mutually enhanced by utilizing the synergistic effect between the tasks; a two-way context network model is adopted as a text recognition model, feature expressions for different texts are fused through a dynamic routing mechanism, two-way time sequence modeling and visual semantic alignment are combined, context dependence of character sequences is considered, and the method is anchored to visual evidences.

The invention relates to a method for realizing high-precision identification of fine inclined character labels for an electrical cabinet pressing plate. The method comprises the following steps of executing operations of color normalization, brightness equalization and distortion correction; character area detection is carried out, and a small character label area is automatically positioned; performing multi-scale feature extraction; normal texts and inclined texts are automatically distinguished; performing rotation, affine or perspective transformation; performing digital full-angle to half-angle conversion, character misrecognition correction, fuzzy word normalization, semantic rule replacement and context consistency verification on the output recognition result; and carrying out structure optimization on the trained identification model. According to the method and the device for realizing high-precision identification of the fine inclined character label for the electrical cabinet pressing plate, the processor and the computer readable storage medium thereof, by combining text type classification, geometric correction and a model acceleration optimization strategy, the identification precision, the robustness and the deployment efficiency of a system in a complex industrial scene are remarkably improved.

The invention belongs to the technical field of digital pathology and artificial intelligence, and particularly relates to an intelligent pathological specimen classification and recognition system based on artificial intelligence. Comprising an image acquisition module, an image preprocessing module, a feature extraction module, a feature fusion module, a classification identification module, a classification modeling module, a lesion area positioning module, a lesion area proportion calculation module, a comprehensive diagnosis scoring module and a result output module. Through a color normalization formula, image differences caused by different dyeing conditions and scanning equipment are reduced, and the model stability is improved; through calculation of a pathology classification index CI, quantification and interpretability of a pathology classification process are realized; automatic positioning and quantitative evaluation of the lesion area are realized through calculation of the lesion probability and the lesion area proportion; the classification result and lesion area information are fused through comprehensive diagnosis scores, so that the accuracy of pathological auxiliary diagnosis is improved; the whole technical scheme does not depend on a specific neural network structure, engineering implementation is flexible, and the protection range is reasonable.

This invention discloses a method for sorting fresh tea leaves based on a frequency-domain tree-structured topology network, along with a computer device and storage medium. The method encompasses a complete process: image acquisition, preprocessing, frequency-domain decomposition, depth modeling, map construction, and classification. First, image quality is improved through color normalization and edge enhancement. Then, frequency-domain tree-structured decomposition using wavelet transform and discrete cosine transform is performed to extract multi-scale features. Subsequently, long- and short-range dependency modeling and residual convolution modules are integrated to achieve multi-level feature representation. Finally, a tree-structured topology attention path and structure map simulating the bud-leaf-vein relationship are constructed to enhance semantic understanding. A tree-structure-aware classification function is used to achieve fine-grained classification of single buds, one bud and one leaf, one bud and two leaves, and one bud and multiple leaves. During training, cross-entropy loss, data augmentation, and regularization strategies are combined to improve model robustness. This method demonstrates high classification accuracy and stability on multiple tea image datasets, effectively improving the practicality of automatic sorting of fresh tea leaves.