1245 results about "Fundus image" patented technology

The invention discloses a fundus image retinal vessel segmentation method and a fundus image retinal vessel segmentation system based on deep learning. The fundus image retinal vessel segmentation method comprises the steps of performing data amplification on a training set, enhancing an image, training a convolutional neural network by using the training set, segmenting the image by using a convolutional neural network segmentation model to obtain a segmentation result, training a random forest classifier by using features of the convolutional neural network, extracting a last layer of convolutional layer output from the convolutional neural network, using the convolutional layer output as input of the random forest classifier for pixel classification to obtain another segmentation result, and fusing the two segmentation results to obtain a final segmentation image. Compared with the traditional vessel segmentation method, the fundus image retinal vessel segmentation method uses the deep convolutional neural network for feature extraction, the extracted features are more sufficient, and the segmentation precision and efficiency are higher.

Systems and methods of obtaining and recording fundus images by minimally trained persons, which includes a camera for obtaining images of a fundus of a subject's eye, in combination with mathematical methods to assign real time image quality classification to the images obtained based upon a set of criteria. The classified images will be further processed if the classified images are of sufficient image quality for clinical interpretation by machine-coded and / or human-based methods. Such systems and methods can thus automatically determine whether the quality of a retinal image is sufficient for computer-based eye disease screening. The system integrates global histogram features, textural features, and vessel density, as well as a local non-reference perceptual sharpness metric. A partial least square (PLS) classifier is trained to distinguish low quality images from normal quality images.

A local adaptive method is proposed for automatic detection of microaneurysms in a digital ocular fundus image. Multiple subregions of the image are automatically analyzed and adapted to local intensity variation and properties. A priori region and location information about structural features such as vessels, optic disk and hard exudates are incorporated to further improve the detection accuracy. The method effectively improves the specificity of microaneurysms detection, without sacrificing sensitivity. The method may be used in automatic level-one grading of diabetic retinopathy screening.

An apparatus for obtaining a retinal image of an eye has a control logic processor (214) for executing a sequence of operations for obtaining the image. A visual target (162) orients the eye of a patient when viewed. An indicator element (410) provides a signal that indicates that the patient is in position. A cornea focus detection section (450) indicates cornea focus, in cooperation with the control logic processor (214). An alignment actuator aligns the optical path according to a signal obtained from the cornea focus detection section (450). A retina focus detection section (452) detects retina focus in cooperation with the control logic processor (214). A focusing actuator (406) is controlled by instructions from the control logic processor (214) according to a signal obtained from the retina focus detection section (452). An image capture light source is energized by the control logic processor (214) for illuminating the retina.

The invention discloses an eye fundus image characteristics extraction method for diabetic retinopathy by utilizing an eye fundus image shot by a digital mydriatic-free eye fundus camera. The method comprises the following steps of (1) selecting an RGB channel of the eye fundus image; (2) positioning the optic disk of the eye fundus image; (3) extracting the hard exudate characteristics and the cotton wool spot characteristics of the eye fundus image, generating an extracted eye fundus image if at least one characteristic is discovered, if no characteristic is discovered, extracting microaneurysm characteristics and retinal hemorrhage characteristics, and generating an extracted eye fundus image. The method belongs to a non-intrusive technology, and is simple, rapid, convenient and effective, and the processed image is clear and obvious in symptom, so the diagnosis of a doctor is facilitated.

To provide a technology that makes it easy to capture a positional relation of a 2-dimensional image of the surface of a fundus oculi and a tomographic image of the fundus oculi. The fundus observation device 1 is provided with a fundus camera unit 1A for forming a 2-dimensional image of the surface of a fundus oculi Ef of an eye to be examined, an OCT unit 150 for forming a tomographic image of the fundus oculi Ef, a scanning unit 141 as well as an image processing part 220, a display 207, and a controlling part 210 of a computer 200 for displaying the 2-dimensional image formed by the fundus camera unit 1A and a tomographic image formed by the OCT unit, etc. in parallel on the display 207, and also for displaying the cross-sectional position information (attention line L, attention region P) indicating the cross-sectional position of the tomographic image on the surface of the fundus oculi Ef, so as to be overlapped with the 2-dimensional image.

The present invention provides a fundus image classification method and device of retinopathy. The method comprises a step of obtaining the fundus image of an eye detection patient, a step of preprocessing the fundus image and obtaining a corresponding grey fundus image, a step of extracting a preset image characteristic in the grey fundus image, a step of analyzing the preset image characteristic according to the diabetic retinopathy classification model of the fundus image, and obtaining the classification result of the corresponding fundus image, a step of extracting the preset image characteristic in the preprocessed grey fundus image, and carrying out prediction classification on the above preset image characteristic through the diabetic retinopathy classification model established according to a support vector machine and a deep neural network so as to obtain a final classification result. The artificial participation is not needed in the whole process, the workload of doctors is reduced, the high standard requirement of the medical qualification level of the doctor is reduced, the accuracy of classification is improved, and the method and the device have great significance for specific applications and medical research.

A system and method for fundus imaging wherein selective illumination of a sector of the field of view of the fundus, using off-axis illumination, provides images of acceptable clarity, resolution, and size, with significantly reduced reflections, in a compact system. By rotating illumination around the optical axis, sectors of the fundus may be selectively and sequentially illuminated. An image of the entire field of view of the fundus is obtained by combining images, e.g. two or more half images, obtained within a single shutter exposure or capture period. An illumination system using LED light sources and a rotatable occluder provides for a lightweight, handheld and portable fundus imaging system. It may take the form of a fundus camera, a fundus imaging lens module for regular camera, or an adapter which couples to the standard lens of a camera to create a low cost fundus camera.

The implementation way of the invention provides a head-mounted display device and an image adjusting method. The head-mounted display device comprises an eyesight detection module, an image adjusting module and a display module, wherein the eyesight detection module is used for shooting fundus imaging of a user, transmitting an image adjusting command to the image adjusting module on the basis of the shot fundus imaging stat and determining the optimal focusing parameter of the fundus imaging of the user in the dynamic adjustment process of the display attribute of the display module; the image adjusting module is used for dynamically adjusting the display attribute of the display module according to the image adjusting command; and the display module is used for receiving an image signal and displaying the image signal to the user on the basis of the display attribute when the optimal focusing parameter exists. According to the implementation way of the invention, the eyesight condition of the user can be identified automatically, so that the user can see the display image clearly and convenience and usability are improved; and the users with different eyesight can use the implementation way of the invention conveniently.

In the context of the early detection and monitoring of eye diseases, such as glaucoma and diabetic retinopathy, the use of optical coherence tomography presents the difficulty, with respect to blood vessel segmentation, of weak visibility of vessel pattern in the OCT fundus image. To address this problem, a boosting learning approach uses three-dimensional (3D) information to effect automated segmentation of retinal blood vessels. The automated blood vessel segmentation technique described herein is based on 3D spectral domain OCT and provides accurate vessel pattern for clinical analysis, for retinal image registration, and for early diagnosis and monitoring of the progression of glaucoma and other retinal diseases. The technique employs a machine learning algorithm to identify blood vessel automatically in 3D OCT image, in a manner that does not rely on retinal layer segmentation.

The object of the invention is to provide a favorable spectral characteristic that reduces variation depending on the frequency of received light intensity, and that is gentle on a subject eye. It also eliminates displacement between positions of respective spectral images of the same part even if a change in alignment occurs between the eye and apparatus with the lapse of time. An apparatus 1 for measuring spectral fundus image data of this invention comprises: an illumination optical system 10 having an illumination light source 11 that emits a light beam in a specified wavelength range; a light receiving optical system 20 for forming a fundus image on the light receiving surface of a photographing section 4; a liquid crystal wavelength tunable filter 32 capable of choosing a wavelength of a transmitted light beam in a specified wavelength range; a spectral characteristic correction filter 13 having wavelength characteristic for correcting the wavelength characteristic of the emitted light intensity of the illumination light source 11 and the transmission wavelength characteristic of the wavelength tunable filter 32 so that the received light intensity on the light receiving surface is kept within the specified range; and a data measuring section 7 for taking the spectral fundus image data from the light receiving surface while changing the wavelength of the light beam passing through the wavelength tunable filter 32.

The invention discloses a vessel-based automatic registration method for an eye fundus image and a spectral-domain optical coherence tomography (SD-OCT) projection image, belonging to the technical field of image processing. The method comprises the following steps: performing tissue layer segmentation on an input SD-OCT image, locating the upper boundary of a photoreceptor inner-segment / outer segment (IS-OS) layer and the lower boundary of a retinal pigment epithelium (RPE), generating an SD-OCT projection image according to the upper boundary of the IS-OS layer and the lower boundary of an RPE, and segmenting the SD-OCT projection image by using an improved level set method; segmenting a colored eye fundus image by using a morphology multi-scale top hat method; searching for the local maximum value of a maximum similar function by using an iteration blocking method to extract a control point, and calculating an affine variation coefficient according to the control point to realize registration of the SD-OCT projection vessel image and the colored eye fundus image. As proved by an experimental result, the image registration accuracy is high in different modes.

The application discloses a depth-learning-based eye-fundus image processing method, device and system. The method comprises: eye-fundus image pretreatment is carried out and region segmentation extraction is carried out on the eye-fundus image; resampling is carried out on a regional image obtained by the segmentation extraction; data expansion is carried out on the regional image after resampling; and a regional image feature is identified by using a depth learning method. Therefore, an eye-fundus image can be analyzed automatically; the analysis result is objective and accurate; and the manpower cost is saved.

A second retinal characterization data set is mapped to a first retinal characterization dataset. The first retinal characterization dataset is displayed as a first graphical map. The second retinal characterization dataset is displayed as a second graphical map which is mapped to the first graphical map. The second graphical map may be warped and morphed onto the first graphical map. Retinal characterization datasets may be derived either from a fundus image or from a retinal parameter dataset calculated from a three-dimensional optical coherence tomography scan of a retina. Retinal parameter datasets may characterize parameters such as retinal thickness. In an embodiment, a fundus image is warped and morphed onto a retinal surface topographical map.

A fundus camera, in which operability is improved while not impairing the convenience of the function of automatic alignment, has a photographing part including a fundus photographing optical system, a moving unit which moves the photographing part, an observation optical system having a first image-pickup element which picks up a fundus image, an observation optical system having a second image-pickup element which picks up an anterior-segment image, a display unit capable of displaying the images, a control part which obtains information on alignment of the photographing part with the eye and controls driving of the moving unit based on the alignment information so that an alignment state satisfies a first reference condition, and a first input switch for inputting a signal for display switching from the fundus image to the anterior-segment image, wherein the control part prohibits the driving of the moving unit when the signal is inputted.

Disclosed are method and apparatus for mapping retinal thickness values to a movable measurement grid. A three-dimensional volume dataset acquired from three-dimensional optical coherence tomography is registered to a fundus image by rendering a two-dimensional composite image from the three-dimensional volume dataset and superimposing characteristic features in the two-dimensional composite image upon corresponding characteristic features in the fundus image. A measurement grid is displayed on the two-dimensional composite image. The measurement grid is moved to a region of interest, and retinal thickness values in the region of interest are mapped to sectors within the measurement grid.

The invention discloses an image detection and processing method and device and a terminal. The image detection method comprises the steps that any one or any combination of stain / bright spot detection, exposure detection, definition detection, light leakage detection and local shadow detection is performed on multiple images after multiple images are acquired. The above mentioned detection can beperformed on the eye fundus image in turn, partial detection can also be performed according to the actual situation and omnidirectional detection of the image can also be realized so that the photographing quality of the image can be effectively detected.

An ocular fundus imager (8) automatically aligns fundus illuminating rays to enter the pupil (P) and to prevent corneal reflections from obscuring the fundus image produced. Focusing the produced fundus image is automatically performed and is based upon the fundus image iself. A head restraint for the patient undergoing examination is in the form of a pair of spectacles which is not only easy to use accurately but significantly reduce the gross alignment between the optical system (8) and the patent's pupil (P).

The present invention relates to the design and implementation of a three stage computer-aided screening system that analyzes fundus images with varying illumination and fields of view, and generates a severity grade for diabetic retinopathy (DR) using machine learning. In the first stage, bright and red regions are extracted from the fundus image. An optic disc has similar structural appearance as bright lesions, and the blood vessel regions have similar pixel intensity properties as the red lesions. Hence, the region corresponding to the optic disc is removed from the bright regions and the regions corresponding to the blood vessels are removed from the red regions. This leads to an image containing bright candidate regions and another image containing red candidate regions. In the second stage, the bright and red candidate regions are subjected to two-step hierarchical classification. In the first step, bright and red lesion regions are separated from non-lesion regions. In the second step, the classified bright lesion regions are further classified as hard exudates or cotton-wool spots, while the classified red lesion regions are further classified as hemorrhages and micro-aneurysms. In the third stage, the numbers of bright and red lesions per image are combined to generate a DR severity grade. Such a system will help in reducing the number of patients requiring manual assessment, and will be critical in prioritizing eye-care delivery measures for patients with highest DR severity.

The invention discloses an optic disc projection location method synthesizing vascular distribution with video disc appearance characteristics, which comprises the following steps of: (1) extracting an interest retina fundus image region by use of a mask operation; (2) carrying out normalization enhancement on a fundus image based on an image observation model; (3) achieving extraction and segmentation of the fundus image by using a non-vascular structure inhibition operator and combining with a hysteresis multi-threshold processing technology; (4) setting a vertical window of double main blood vessel width to slide on a blood vessel segmentation image along the horizontal direction and calculating a vascular distribution degree D(x) at each horizontal position x to obtain a distribution degree curve of a horizontal projection, wherein the minimum value point of the curve is determined to be optic disc horizontal coordinate xod; (5) setting a rectangular window side length of which is equal to the optic disc diameter to slide up and down at the horizontal coordinate xod, estimating local region brightness IN(xod, y) and edge gradient information gN(xod, y) respectively, drawing a vertical projection curve reflecting a change of a characteristic value f(y)=IN(xod, y)*gN(xod, y), wherein a maximum value point of the curve is optic disc vertical coordinate yod. The method has a simple algorithm, a high success rate and excellent robustness.

An ophthalmologic photographing apparatus includes a light source having two or more types of light-emitting-diode (LED) light emitting elements configured to emit light of different wavelengths, an illumination optical system configured to illuminate a fundus of a subject's eye with light emitted by the light source, an imaging optical system configured to form a fundus image from light reflected from the fundus of the subject's eye illuminated by the illumination optical system, and a control unit configured to control the light source to emit light of different wavelengths, with which the fundus of the subject's eye is illuminated, between when the formed fundus image is observed and when the formed fundus image is captured as a still image.

The invention discloses a fundus scanning imaging device which is used for fundus examination of a human eye medically. A human eye fundus image is obtained through a confocal scanning technology to judge the fundus pathology and focus. The fundus scanning imaging device adopts laser as a lighting source. A scanning galvanometer achieves two-dimensional scanning of the human eye fundus. A high-sensitivity photodiode is used for detecting light signals returned from the human eye fundus, the collected signals are subjected to signal conversion and image reconstruction to obtain a large-view-field high-resolution fundus retina video image. The fundus scanning imaging device adopts a single galvanometer to replace a traditional orthogonal double-galvanometer module and acquires high-frame-frequency video image outputting through low galvanometer scanning frequency. A single detector is adopted, and a motor is utilized to switch an optical filter to achieve an FFA and ICGA fluorescent radiography function. Two kinds of retina images of light sources with different wavelengths can be utilized to combine a pseudo color image by synchronously controlling the image collection sequence, and two functions of video and camera shooting can be achieved simultaneously.

The invention provides an accurate and reliable method and device for processing eye fundus images based on deep learning in order to resolve problems in the prior art. The method for processing eye fundus images based on deep learning comprises the following steps: obtaining multiple original sample eye fundus images and corresponding lesion grade parameters; performing image pre-processing operation and image data enhancing operation on the multiple original sample eye fundus images in order to obtain multiple training sample eye fundus images; performing deep-network training on the multiple original sample eye fundus images and the corresponding lesion grade parameters in order to obtain an eye fundus lesion grade recognition model; and inputting test eye fundus images into the eye fundus lesion grade recognition model in order to obtain lesion grade recognition results.

A method of processing an image includes processing the image so as to produce a modified image in which the width of each line object is a single pixel, identifying pixels of the modified image which are relocated on a line object, and allocating to each image point a score value dependent upon the number of adjacent pixels which are likewise located on a line, determining from the score values which of the image points is disassociated with a crossing point or a bifurcation of the respective line object, performing a matching operation on pairs of line segments for each crossing point, classifying the line objects in the original image into two arbitrary sets, and designating one of the sets as representing venous structure, the other of the sets as representing arterial structure.

A model eye that models the optical characteristics of a human eye and is endowed with a grayscale pattern on the ocular fundus model surface is stereographically photographed with a parallax via a stereo photographic optical system. Photographed images are processed to provide calibration data for correcting the shape distortions of the stereo photographic optical system. The calibration data is used to correct a distortion-affected shape data and parallax images obtained in stereographic photography of the actual ocular fundus of a subject's eye. The shape distortion-corrected parallax image is used for a three-dimensional measurement process and 3D display on a stereo monitor. This allows an accurate three-dimensional measurement to be carried out and an accurate fundus image to be produced. The examiner can accurately evaluate the stereo shape of the ocular fundus of the subject's eye.

To accurately detect a movement of an object based on an image distorted by the movement of an eye to be inspected, which is acquired by a scanning imaging system, provided is an ophthalmologic apparatus including: an extraction means for extracting a plurality of characteristic images from a first fundus image of the eye to be inspected; a fundus image acquisition means for acquiring a second fundus image of the eye to be inspected during a period different from a period during which the first fundus image is acquired; and a calculation means for calculating at least a rotation of movements of the eye to be inspected based on the plurality of characteristic images and the second fundus image.

A system and computer-implemented method for detecting retinopathy is provided. The system comprises an image input module configured to receive one or more fundus images. Further, the system comprises a pre-processing module configured to apply one or more transformations to the one or more received fundus images. Furthermore, the system comprises a feature extraction module configured to extract one or more features from the one or more transformed images using one or more Convolutional Neural Networks (CNNs). Also, the system comprises a prediction module configured to determine stage of retinopathy by classifying the one or more extracted features using pre-stored features, wherein the pre-stored features are extracted from one or more training fundus images by the one or more CNNs and further wherein each pre-stored feature corresponds to a class which is associated with a predetermined stage of retinopathy.