Abstract: A capsule endoscope image recognition method based on deep learning, and a device and a medium is provided. After a plurality of frames of continuously captured images are processed to form an image sequence of a specific format, multi-channel recognition is performed on the plurality of frames of images by means of a 3D convolutional neural network model, and a recognition result is then output in combination with a recognition probability of each channel, such that the image recognition precision is improved.

Abstract: The present invention provides a three-dimensional reconstruction method of a capsule endoscope image, an electronic device, and a readable storage medium, which combines two algorithms for obtaining image depth information values to improve the calculation accuracy of image depth information values, and thus to increase the image three-dimensional reconstruction rate and improve the image identification precision.

Abstract: A referenceless image evaluation method for a capsule endoscope, an electronic device, and a medium. The method comprises: acquiring image quality evaluation scores and image content evaluation scores corresponding to respective original images; and determining, according to weights of the image content evaluation scores and weights of the image quality evaluation scores, a comprehensive score for each image currently being evaluated. The referenceless image evaluation method for a capsule endoscope employs various evaluation models to evaluate the image quality and content of a number of original images taken in the same detection region. The results of these evaluations are then combined to produce comprehensive scores for the numerous original images of the same region. Through the comprehensive scores, high-quality images can be quickly screened out, and thus the screening of the original images can be accelerated and the identification accuracy be improved.

Abstract: A method for stitching images of a capsule endoscope, an electronic device, and a readable storage medium are provided. The method comprises: performing image rectification and circular edge masking on original images to form pre-processed images, and performing image enhancement (S1); completing detection and pairing of feature points (S2); calculating a transformation model of all pre-processed images to a same optimal plane according to a set of the feature points (S3); performing projective transformation for each enhanced image to a same coordinate system (S4); and stitching images according to an obtained sequence to form a fused image for output (S5), thus expanding the field of view of a single image of the capsule endoscope.

Abstract: The present invention discloses a capsule endoscope system, a method of identifying stained area in endoscopic images, and a computer readable storage medium. The method comprises: obtaining a basic image taken by a photographing device; removing abnormal pixels from the basic image and recording it as a processed image; obtaining the hue, saturation and value of each pixel in the processed image of HSV format; setting a first condition that the hue is in a D1 range, the saturation is in a D2 range, and the value is in a D3 range; obtaining the b component of each pixel in the processed image of Lab format; setting a second condition that the b component is in a D4 range; collecting and processing the pixels that meet the first condition and/or the second condition to form a set C; calculating the initial range area S1 of the set C.

Abstract: The present invention provides an image-based position detection method, an electronic device, and a readable storage medium. The method includes: driving a capsule endoscope to continuously capture a group of images at an interval of a first preset time t1, where each group of images comprises at least one image; obtaining at least one image from each group of images, and calculating a determination index according to grayscale values of at least two channels of the image; determining whether the capsule endoscope reaches a preset detection position according to a value of the determination index; if yes, adjusting the capsule endoscope to a corresponding working mode according to a current position of the capsule endoscope, and if no, continue monitoring. The present invention can determine whether the capsule endoscope reaches the preset detection position, which greatly save battery power, and improve work efficiency of the capsule endoscope.

Abstract: A method for stitching images of a capsule endoscope, an electronic device, and a readable storage medium are provided. The method comprises: performing image rectification and circular edge masking on original images to form pre-processed images, and performing image enhancement (S1); completing detection and pairing of feature points (S2); calculating a transformation model of all pre-processed images to a same optimal plane according to a set of the feature points (S3); performing projective transformation for each enhanced image to a same coordinate system (S4); and stitching images according to an obtained sequence to form a fused image for output (S5), thus expanding the field of view of a single image of the capsule endoscope.

Abstract: The present invention provides a transfer learning based capsule endoscopic images classification system. The system removes the capsule endoscopic images with an average brightness value beyond the preset threshold, and removes the capsule endoscopic images without details based on image brightness standard deviation and image brightness gradient. The system also removes similar images from the capsule endoscopic images using optical flow method, classifies the capsule endoscopic images according to the corresponding anatomical structure, and obtains the classified capsule endoscopic images list arranged in chronological order. The system further determines and labels the position of the first image of each specific anatomical structure in the classified capsule endoscopic images list arranged in chronological order.

Abstract: The present invention discloses a capsule endoscope system, a method of identifying stained area in endoscopic images, and a computer readable storage medium. The method comprises: obtaining a basic image taken by a photographing device; removing abnormal pixels from the basic image and recording it as a processed image; obtaining the hue, saturation and value of each pixel in the processed image of HSV format; setting a first condition that the hue is in a D1 range, the saturation is in a D2 range, and the value is in a D3 range; obtaining the b component of each pixel in the processed image of Lab format; setting a second condition that the b component is in a D4 range; collecting and processing the pixels that meet the first condition and/or the second condition to form a set C; calculating the initial range area S1 of the set C.

Abstract: The present invention discloses a method, device and medium for structuring a capsule endoscopy report text. The method includes: annotating the report text using an annotation model; storing each named entity classification label in the report text in a hierarchical tree structure according to the annotation information to form a tree structure diagram; parsing the tree structure diagram, extracting abnormal structure data and time parameters, and storing the abnormal structure data and time parameters in a panel data structured manner to form an abnormal structure panel table and a time parameter panel table. The present invention can automatically annotate the capsule endoscopy report through the annotation model, and output parameters of different amount of information in different structures, and quantitative and accurate quality control of the capsule endoscopy process and examination results, which provide sufficient convenience for the electronic medical information of the capsule endoscopy.

Abstract: The present invention discloses a method, an apparatus and a storage medium for labeling capsule endoscopy report. The method includes: collecting p report samples to establish an initial corpus database; parsing the report samples in the initial corpus database, to establish a named entity recognition dictionary and a pattern rules database, and removing duplicate texts from the named entity recognition dictionary and the pattern rules database; since the q-th report sample is collected, q=p+1, querying the named entity recognition dictionary and pattern rules database with texts appearing in the report sample, to automatically label the current report sample. The present invention can build a database by parsing a small number of labeled report samples, make subsequent report samples query the database using specific rules, and then label the report samples automatically in a fast and effective manner, save labor costs and improve labelling efficiency.

Abstract: An auxiliary display system for a photographing device is provided. The photographing device takes images in vivo and generates image information. The auxiliary display system includes a server and a client in communication with the server. The server sequentially identifies and analyzes the image information in real time and generates suspected lesion information. The client receives suspected lesion information transmitted from the server, obtains image information corresponding to the suspected lesion information, and simultaneously displays the corresponding image information and the suspected lesion information in real time.

Abstract: The present invention provides a transfer learning based capsule endoscopic images classification system. The system removes the capsule endoscopic images with an average brightness value beyond the preset threshold, and removes the capsule endoscopic images without details based on image brightness standard deviation and image brightness gradient. The system also removes similar images from the capsule endoscopic images using optical flow method, classifies the capsule endoscopic images according to the corresponding anatomical structure, and obtains the classified capsule endoscopic images list arranged in chronological order. The system further determines and labels the position of the first image of each specific anatomical structure in the classified capsule endoscopic images list arranged in chronological order.

Abstract: A system and a method for preprocessing capsule endoscope images are provided. The capsule endoscopic image preprocessing system include an in vitro image removal module, an invalid image removal module, a digestive tract image classification module, a lesion and anatomical structure identification module, and a lesion and anatomical structure redundant image removal module. The capsule endoscopic image preprocessing system removes in vitro images and invalid images from capsule endoscopic images, classifies the capsule endoscopic images according to different parts of the digestive tract, identifies lesion and anatomical structures in the classified capsule endoscopic images; and removes redundant lesion and anatomical structure images according to the lesion and anatomical structures.

Abstract: A system and a method for preprocessing capsule endoscope images are provided. The capsule endoscopic image preprocessing system include an in vitro image removal module, an invalid image removal module, a digestive tract image classification module, a lesion and anatomical structure identification module, and a lesion and anatomical structure redundant image removal module. The capsule endoscopic image preprocessing system removes in vitro images and invalid images from capsule endoscopic images, classifies the capsule endoscopic images according to different parts of the digestive tract, identifies lesion and anatomical structures in the classified capsule endoscopic images; and removes redundant lesion and anatomical structure images according to the lesion and anatomical structures.