11 results about "Cervical cell" patented technology

The application relates to the technical field of medical image processing, and particularly discloses a cervical cancer screening pathological large model construction method based on DINOv3-TCT. The technical problem to be solved by the application is that when a general visual Transformer model is directly applied to a cervical cell instance segmentation task, there are problems of mismatch between model internal feature representation and instantiation target and high calculation redundancy. Therefore, the application proposes to transform the DINOv3 encoder: inject a learnable query token at a specific layer, and innovatively construct an instance decoupling loss term acting on an attention weight matrix to explicitly guide the encoder to realize instance-level feature separation during training. The method optimizes the internal attention distribution of the model, improves the accuracy of dense cell segmentation, and simultaneously reduces the model parameter quantity and calculation overhead through integrated design, so that the balance between high accuracy and high efficiency is realized.

1. The name of the design product: brush head for collecting cervical cells. 2. The use of the design product: the product is a device for collecting cervical exfoliative cells of women, mainly used for Papanicolaou smear (PAP) test. 3. The design points of the design product: in shape. 4. The picture or photo that best indicates the design points: perspective view 1.

The application discloses a kind of based on position perception and feature modulation cell image multi-task classification method and system, belong to medical image analysis field.The method first utilizes the global visual feature of cervical cell image extracted by depth backbone network, and predicts its hierarchical position (surface layer / middle layer / base layer).On this basis, introduce position perception hybrid expert module, utilize position information by two paths: one will position label be mapped into semantic embedding feature;Another generates channel level modulation parameter according to position label, carries out dynamic affine transformation to global feature, and the feature after modulation is sent into multiple expert networks.At the same time, based on the routing network of image content generates dynamic weight, weights the fusion of expert feature, obtains enhanced feature.Finally, enhanced feature and position semantic embedding feature are fused, and lesion type classification and position classification tasks are collaboratively trained.The application guides feature modulation and expert fusion by position information, effectively improves the accuracy and explainability of cervical cell image classification, and can be used for computer-aided screening of cervical cancer.

The present application belongs to the technical field of slice image recognition, and particularly relates to a cervical cell pathological slice recognition method based on multi-modal learning, which comprises the following steps: acquiring cervical pathological slice images and clinical text data of a patient, extracting image modal feature vectors and text modal feature vectors through an image encoder and a text encoder respectively; inputting the two modal features into a cross-modal fusion network constructed based on an asymmetric attention mechanism, guiding image feature enhancement with the text feature as a query, guiding text feature enhancement with the image feature as a query, and fusing to generate multi-modal joint feature representation; and outputting classification results such as normal cells, low-grade lesions, high-grade lesions or squamous cell carcinoma based on the joint feature representation. Through bidirectional cross-modal attention interaction, the present application realizes effective fusion of pathological image morphological features and clinical text information, and makes up for the limitations of single-modal data.

The application provides a cervical cell nucleus multi-scale accurate segmentation method based on an HSV channel difference feature, the method first converts an image to an HSV space and extracts a saturation S and a brightness V channel; then, adaptive enhanced difference feature channels are constructed by fusing a weighted difference value and a ratio relationship, the discrimination of the cell nucleus and the background is effectively improved, and the method has strong robustness to staining and illumination changes; subsequently, cell density sensing adaptive CLAHE and gradient guided adaptive median filtering are used for multi-scale enhancement and denoising, interference is suppressed while details are retained; in the segmentation stage, a hierarchical strategy of a global threshold value, a local adaptive threshold value and a morphological gradient enhancement is adopted, and adaptive morphological post-processing and screening are combined, so that accurate and complete segmentation of the cell nucleus in a complex scene is finally realized; the whole process of the application is based on clear image processing principles, does not require large-scale training data, has high calculation efficiency and is easy to integrate into an existing pathological information system.

The application provides an immunochemical luminescence kit for detecting a cancer marker HPV18 E7 protein, the kit comprising biotin-labeled antibody 1 and acridinium ester-labeled antibody 2; the antibody 1 is a rabbit-derived monoclonal antibody, comprising a heavy chain and a light chain; the antibody 1 heavy chain variable region comprises VHCDR1, VHCDR2 and VHCDR3 with the amino acid sequences shown in SEQ ID NO. 1-3; the antibody 1 light chain variable region comprises VLCDR1, VLCDR2 and VLCDR3 with the amino acid sequences shown in SEQ ID NO. 9-11; the antibody 2 is a rabbit-derived monoclonal antibody, comprising a heavy chain and a light chain; the antibody 2 heavy chain variable region comprises VHCDR1, VHCDR2 and VHCDR3 with the amino acid sequences shown in SEQ ID NO. 17-19; and the antibody 2 light chain variable region comprises VLCDR1, VLCDR2 and VLCDR3 with the amino acid sequences shown in SEQ ID NO. 25-27. The application can improve the test sensitivity, improve the detection rate of weak positive samples, realize accurate quantification of HPV18 E7 protein in cervical exfoliated cells, and the detection method of the application is simple in operation, fully automatic in the detection process, and can greatly improve the detection efficiency.

The application discloses a cervical cell seven-classification method and system based on hierarchical routing and boundary expert fusion, and belongs to the technical field of image processing. In order to solve the technical problem that the existing technology directly adopts a single multi-classification model for cell classification, resulting in unreliable gray area boundary discrimination and finally leading to unstable cell classification result discrimination, the application utilizes multi-dimensional morphological indexes and detection confidence to calculate quality points; single cells and cell groups and halo cells are shunted for processing, the cell groups and the halo cells are directly outputted in types, the single cells enter subsequent processes, different morphological cells are avoided from being mixed into the same classifier, the single model training target inconsistency and feature deviation problem is solved; the single cell five-classification basic probability is acquired firstly, target areas of ASC-US and LSIL and ASC-H and HSIL are screened out through boundary uncertainty comprehensive points and quality points, finally probability is obtained through fusion of morphological indexes, bias probability and fusion strength, and the class with the maximum probability is selected as the seven-classification result. The application is used for cervical cell classification.

This invention belongs to the field of image classification technology, specifically involving a cervical cell pathology slide classification method based on weakly supervised learning. First, key image patches are screened using a dual index of feature entropy and activation heat, and mapped to instance feature sets. The teacher branch calculates instance attention weights based on package-level labels, outputs package-level predictions after weighted aggregation, and generates soft pseudo-labels by normalizing the weights. The student branch fits the distribution of soft pseudo-labels through knowledge distillation and generates hard pseudo-labels. After fusing distillation and cross-entropy loss, the shared encoder parameters are updated. Finally, the updated encoder parameters are synchronized to the feature extraction stage. The model is iteratively optimized through alternating training by the teacher and student branches and a difficult instance mining mechanism, outputting classification results and generating a heatmap of positive instance location. This invention achieves high-precision instance-level classification and positive region location under weak supervision using only slide-level labels, effectively improving the identification ability of difficult positive instances.

The application provides an immunochemiluminescence kit for detecting a cancer marker HPV16 E7 protein, the kit comprising biotin-labeled antibody 1 and acridinium ester-labeled antibody 2; the antibody 1 is a mouse monoclonal antibody, comprising a heavy chain and a light chain; the amino acid sequence of the antibody 1 heavy chain is shown in SEQ ID NO: 17, and the amino acid sequence of the antibody 1 light chain is shown in SEQ ID NO: 18; the antibody 2 is a rabbit-derived monoclonal antibody, comprising a heavy chain and a light chain; the antibody 2 heavy chain variable region comprises VHCDR1, VHCDR2 and VHCDR3 with the amino acid sequences shown in SEQ ID NO. 1-3; and the antibody 2 light chain variable region comprises VLCDR1, VLCDR2 and VLCDR3 with the amino acid sequences shown in SEQ ID NO. 9-11. The application can improve test sensitivity, improve the detection rate of weak positive samples, identify early cervical cancer, realize accurate quantification of HPV16 E7 protein in cervical exfoliated cells, and has the advantages of simple operation, automatic detection process and greatly improved detection efficiency.

This invention provides an automated screening method and system for cervical cell tumors. The method includes: dividing the foreground region of a pathological slide image to be screened into multiple image blocks; extracting the feature vector of each image block; inputting the feature vector of each image block into a second model; inferring the positive probability of each image block through the second model; based on the positive probability of each image block, selecting the K image blocks with the highest positive probability from the pathological slide image to be screened, where K is a positive integer; inputting the feature vectors of the K image blocks into a third model, and outputting the positive probability of the pathological slide image to be screened. This invention selects the K image blocks with the highest positive probability through a second model, then focuses on these K image blocks, and infers the positive or negative status of the entire image block based on the image blocks with the highest positive probability, resulting in more accurate inference results.

This invention discloses a lightweight cervical cell classification method based on multi-scale entropy-guided feature adaptation, belonging to the fields of medical image processing and deep learning technology. The method includes: acquiring and preprocessing cervical cell images; calculating the local entropy of the image using neighborhood windows of three scales (5×5, 7×7, and 9×9) to generate a multi-scale entropy map, which is then weighted and fused to obtain a fused entropy map; concatenating the fused entropy map with the original RGB three-channel input to form a four-channel input tensor; mapping the four-channel input to a three-channel adaptive feature map using a lightweight entropy feature adaptation module consisting of two 1×1 convolutional layers to adapt to a pre-trained MobileNetV2 network; inputting the adaptive feature map into MobileNetV2 for feature extraction and classification, outputting the cell category. This invention introduces an explicit multi-scale texture prior at the data input level, adding only about 0.38M parameters, effectively improving the accuracy and interpretability of cervical cell classification while maintaining lightweight design.