CN112541907A - Image identification method, device, server and medium - Google Patents
Image identification method, device, server and medium Download PDFInfo
- Publication number
- CN112541907A CN112541907A CN202011501530.6A CN202011501530A CN112541907A CN 112541907 A CN112541907 A CN 112541907A CN 202011501530 A CN202011501530 A CN 202011501530A CN 112541907 A CN112541907 A CN 112541907A
- Authority
- CN
- China
- Prior art keywords
- image
- region
- interest
- size
- contour
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/149—Segmentation; Edge detection involving deformable models, e.g. active contour models
Abstract
The embodiment of the invention discloses an image identification method, an image identification device, a server and a medium. The method comprises the following steps: acquiring an image to be identified, and extracting the outline of an interested area in the image to be identified to generate an outline image; obtaining an interested area image according to the outline image and the image to be identified; and inputting the image of the region of interest into a trained type recognition model, and determining the image type of the image of the region of interest. According to the technical scheme of the embodiment of the invention, the problem that the classification efficiency of the image to be recognized is low because the region of interest in the image to be recognized is recognized manually is solved, the type of the region of interest in the image is determined rapidly, and the diagnosis efficiency of the medical image is improved.
Description
Technical Field
The embodiment of the invention relates to the technical field of image processing, in particular to an image identification method, an image identification device, a server and a medium.
Background
At present, breast cancer becomes a common tumor threatening physical and mental health of women, the early screening is still insufficient, and according to 2009 breast cancer onset data published by the disease prevention and control bureau of the national cancer center and the ministry of health in 2012, the data show that: the incidence of breast cancer in tumor registration areas nationwide is on the 1 st of female malignant tumors.
The mammary gland molybdenum target X-ray photographic examination is the first choice and the simplest and most reliable noninvasive detection means for diagnosing mammary gland diseases. The molybdenum target of the breast can reliably identify benign lesions and malignant tumors of the breast, and is one of the best accepted methods for clinical routine examination and prevention screening of the breast cancer.
At present, the calcification grade of a calcification area in a breast molybdenum target image needs to be judged by a doctor, and the manual judgment efficiency is low.
Disclosure of Invention
The embodiment of the invention provides an image identification method, an image identification device, a server and a medium, which are used for realizing the effects of quickly determining the type of an interested area in an image and improving the diagnosis efficiency of a medical image.
In a first aspect, an embodiment of the present invention provides an image recognition method, where the method includes:
acquiring an image to be identified, and extracting the outline of an interested area in the image to be identified to generate an outline image;
obtaining an interested area image according to the outline image and the image to be identified;
and inputting the image of the region of interest into a trained type recognition model, and determining the image type of the image of the region of interest.
In a second aspect, an embodiment of the present invention further provides an image recognition apparatus, where the apparatus includes:
the contour image acquisition module is used for acquiring an image to be identified and extracting the contour of an interested area in the image to be identified to generate a contour image;
the interesting region image acquisition module is used for obtaining an interesting region image according to the outline image and the image to be identified;
and the image type determining module is used for inputting the image of the region of interest into a trained type recognition model and determining the image type of the image of the region of interest.
In a third aspect, an embodiment of the present invention further provides an image recognition apparatus, where the image recognition apparatus includes:
one or more processors;
storage means for storing one or more programs;
when executed by the one or more processors, cause the one or more processors to implement an image recognition method as provided by any of the embodiments of the invention.
In a fourth aspect, the embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the program, when executed by a processor, implements the image recognition method as provided in any embodiment of the present invention.
According to the technical scheme of the embodiment of the invention, the image to be identified is obtained, and the outline of the region of interest in the image to be identified is extracted to generate the outline image; the image to be recognized is conveniently cut according to the contour image; obtaining an interested area image according to the outline image and the image to be identified; excluding the interference of other area images except the region of interest in the image to be identified on the type identification of the image; the image of the region of interest is input into a trained type recognition model, the image type of the image of the region of interest is determined, the problem that the efficiency is low when the region of interest in the image to be recognized is recognized manually and then the image to be recognized is classified is solved, the type of the region of interest in the image is determined rapidly, and the effect of improving the diagnosis efficiency of medical images is achieved.
Drawings
FIG. 1 is a flowchart of an image recognition method according to a first embodiment of the present invention;
fig. 2 is an exemplary diagram of a contour image in the first embodiment of the present invention.
Fig. 3 is an exemplary diagram of a cutting image in the first embodiment of the invention.
FIG. 4 is a flowchart of an image recognition method according to a second embodiment of the present invention;
FIG. 5 is an exemplary diagram of a target image according to a second embodiment of the present invention;
FIG. 6 is an exemplary view of another target image in the second embodiment of the present invention;
fig. 7 is a structural diagram of an image recognition apparatus according to a third embodiment of the present invention;
fig. 8 is a schematic structural diagram of an image recognition apparatus in a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of an image recognition method, an image recognition apparatus, a server, and a medium according to an embodiment of the present invention, where the embodiment is applicable to a case of recognizing and classifying images, and the method may be executed by the image recognition apparatus, and specifically includes the following steps:
and S110, acquiring an image to be identified, and extracting the outline of the region of interest in the image to be identified to generate an outline image.
Alternatively, the image to be identified may be a medical image having a region of interest, comprising: x-ray images, magnetic resonance images, computed tomography images, molybdenum target images, and the like. The image to be identified also comprises other image information except the region of interest, and the other image information except the region of interest does not help medical diagnosis and sometimes interferes with the medical diagnosis, so that the contour of the region of interest in the image to be identified needs to be extracted, the contour image is generated according to the contour of the region of interest, and the images of other regions except the region of interest in the image to be identified are removed through the contour image, so that the other regions are prevented from interfering with the image identification, and further the medical diagnosis of the image to be identified is influenced.
Acquiring an image to be identified, acquiring a contour of a region of interest in the image to be identified, and optionally, extracting the contour of the region of interest in the image to be identified by adopting an edge extraction method. Or inputting the image to be recognized into a trained contour extraction model to extract the region of interest in the image to be recognized.
Optionally, when the contour extraction model is trained, a sample image of the contour to be extracted is obtained, and the sample image is also a medical image. The method comprises the steps of marking an interested region in an acquired sample image of which the outline is to be extracted to obtain the outline of the interested region, and storing the outline of the interested region and the sample image of which the outline is to be extracted respectively, so that the marked outline of the interested region is prevented from covering image information in the sample image, the complete image information of the sample image cannot be observed later, and medical diagnosis of the sample image is not facilitated. And determining the length and the width of the contour image according to the size of the contour of the region of interest, so that the contour image can contain all the contours of the region of interest, thereby generating a sample contour image, wherein the sample contour image is a binary image, the value of the region of interest in the sample contour image is 1, and the values of the rest regions are 0. And carrying out iterative training on the contour extraction model to be trained based on the sample image of the contour to be extracted and the sample contour image to obtain the contour extraction model.
Optionally, extracting a contour of the region of interest in the image to be identified to generate a contour image includes: extracting the outline of the region of interest in the image to be identified; and generating a contour image according to the size of the contour, wherein in the contour image, the pixel points of the region inside the contour of the region of interest are set to be 1, and the pixel points of the region outside the contour of the region of interest are set to be 0. Extracting the contour of an interested area in an image to be identified, determining the length and the width of a contour image according to the size of the contour, enabling the size of the contour image to be larger than or equal to that of the interested area, generating the contour image, setting pixel points within the contour of the interested area in the contour image as 1, setting pixel points outside the contour of the interested area as 0, enabling the contour image to be a binary image, conveniently segmenting the image to be identified through the contour image, and obtaining the interested area in the image to be identified, wherein the acquired contour image is shown in fig. 2.
And S120, obtaining an image of the region of interest according to the contour image and the image to be identified.
Alternatively, the Hounsfield Unit (HU) value is a measure of how much local tissue or organ density is measured in a human body. The HU value of the image to be recognized is converted into a preset range through the window width and the window level, if the range is 0-255, interference information in the image to be recognized is eliminated, the region of interest of the HU value in the same range is conveniently acquired, the region of interest of the HU value in the same range is input into a trained type recognition model, and the determined type of the region of interest image is more accurate.
Optionally, obtaining an image of a region of interest according to the contour image and the image to be identified, further comprising: carrying out cutting operation on the image to be identified according to the outline image to obtain a cut image, wherein the size of the cut image meets the required size of the image of the region of interest; and multiplying the cutting image and the contour image to obtain the region-of-interest image. And cutting the image to be identified according to the contour image, optionally, the obtained cut image needs to contain all corresponding regions of interest, and the size of the cut image can be set to be the same as that of the contour image. And multiplying the cut image and the contour image to obtain an image of the region of interest. In the region-of-interest image, the pixel value of the region-of-interest is the same as that of the image to be recognized, and the pixel values of the regions other than the region-of-interest are the same as those of the outline image. As shown in fig. 3, the obtained cutting image is obtained, so that the cutting image retains real image information of the region of interest in the region of interest image, and the region of interest image is conveniently classified, so that the classification is more accurate.
And S130, inputting the image of the region of interest into the trained type recognition model, and determining the image type of the image of the region of interest.
And inputting the obtained interested area image into a trained type recognition model to obtain the image type of the interested area image. Optionally, the region-of-interest image includes a lesion region, and the image type may be a type of the lesion region in the region-of-interest image. Optionally, the image to be identified is a medical image of the target object, the region of interest is a calcified region in the medical image, and the image type includes a calcification level of the calcified region. For example, the image to be identified may be a breast molybdenum target image, the region of interest in the breast molybdenum target image is a calcified region, and the calcified region may be subjected to calcification classification, for example, the calcified region is classified into 4 grades, i.e., 1 grade, 2 grade, 3 grade and 4 grade. The calcified regions of grade 1 and grade 2 are classified into benign calcified regions, and the calcified regions of grade 3 and grade 4 are classified into malignant calcified regions, so the image type of the region-of-interest image includes benign or malignant. The images of the region of interest are input into the type identification model, the types of the images of the region of interest are subjected to primary classification judgment, so that patients with malignant lesions can be screened out quickly, the working efficiency of doctors is improved, the medical cost of the patients is reduced, a large amount of physical examination is facilitated, and early screening and early treatment are performed.
According to the technical scheme of the embodiment of the invention, the image to be identified is obtained, and the outline of the region of interest in the image to be identified is extracted to generate the outline image; the image to be recognized is conveniently cut according to the contour image; obtaining an interested area image according to the outline image and the image to be identified; excluding the interference of other area images except the region of interest in the image to be identified on the type identification of the image; the image of the region of interest is input into a trained type recognition model, the image type of the image of the region of interest is determined, the problem that the efficiency is low when the region of interest in the image to be recognized is recognized manually and then the image to be recognized is classified is solved, the type of the region of interest in the image is determined rapidly, and the effect of improving the diagnosis efficiency of medical images is achieved.
Example two
Fig. 4 is a flowchart of an image recognition method, an image recognition apparatus, a server, and a medium according to a second embodiment of the present invention, where this embodiment is further optimized based on the first embodiment, and optionally, before the image of the region of interest is input into the trained type recognition model, the method further includes: converting the region-of-interest image into a target image with a preset size and a preset shape; correspondingly, the inputting the region-of-interest image into the trained type recognition model includes: and inputting the target image into a trained type recognition model. The size and the shape of the image of the region of interest are unified, so that the type recognition model can better classify the image of the region of interest, and the image classification result is more accurate.
As shown in fig. 4, the method specifically includes the following steps:
s210, acquiring an image to be identified, and extracting the outline of the region of interest in the image to be identified to generate an outline image.
And S220, obtaining an image of the region of interest according to the contour image and the image to be identified.
And S230, converting the region-of-interest image into a target image with a preset size and a preset shape.
The size of the region of interest in each image to be recognized is different, so that the size of the obtained region of interest images is different, the region of interest images are converted into preset sizes and preset shapes to generate target images, the obtained target images better meet the requirements of users, the sizes and the shapes of the regions of interest images are unified, a type recognition model is facilitated to classify the region of interest images, and more accurate image types are obtained.
Optionally, converting the region-of-interest image into a target image with a preset size and a preset shape includes: filling pixel points in the interested region image, and converting the interested region image into an intermediate image in the preset shape; and adjusting the size of the intermediate image according to the size of the intermediate image and the preset size to obtain a target image. Because the shape of the region of interest of each image to be recognized is irregular, the obtained outline image is generally a rectangular image, the size of each outline image is different due to the different sizes of the region of interest, the region of interest images obtained through the outline image and the cut image of the image to be recognized are rectangular images with different sizes, and the shape of the region of interest images is converted into a middle image with a preset shape in a pixel filling mode. Illustratively, the size of the long side of the image of the region of interest is determined, pixel filling is performed on the image of the region of interest according to the size of the long side of the image of the region of interest, the image of the region of interest is filled into a square with the long side as the side length, and the image of the region of interest can be filled into an intermediate image in any shape as required.
And filling the image of the region of interest into a preset shape to generate an intermediate image, adjusting the size of the intermediate image, and adjusting the intermediate image to be a preset size, thereby obtaining the target image. Optionally, adjusting the size of the intermediate image according to the size of the intermediate image and the preset size to obtain a target image, including:
determining an interpolation scaling of the intermediate image based on the size of the intermediate image and the preset size, and carrying out size adjustment on the intermediate image based on the interpolation scaling to obtain a target image; or if the size of the intermediate image is smaller than the preset size, performing pixel filling on the intermediate image based on the preset size, and if the size of the intermediate image is larger than the preset size, performing image contraction on the intermediate image based on the preset size to obtain a target image; or inputting the intermediate image and the preset size into an image conversion model, and acquiring a target image which is output by the image conversion model and meets the preset size.
In the size adjustment of the intermediate image, optionally, the size of the intermediate image is compared with a preset size, and an interpolation scaling for converting the size of the intermediate image into the preset size is determined. And according to the interpolation scaling, the intermediate image is enlarged or reduced in an interpolation mode to adjust the size of the intermediate image to be a preset size. Illustratively, the size of the intermediate image is adjusted to the preset size by resize, and fig. 5 shows an image in which the size of the intermediate image is adjusted to the preset size by resize.
Optionally, the size of the intermediate image is compared with a preset size, and if the size of the intermediate image is smaller than the preset size, pixel filling is performed on the intermediate image, so that the size of the filled intermediate image is equal to the preset size, and then the target image is obtained. And when the size of the intermediate image is larger than the preset size, shrinking the intermediate image to enable the size of the shrunk intermediate image to be equal to the preset size, so that the target image is obtained. For example, the size of the intermediate image may be adjusted to a preset size by padding, and as shown in fig. 6, the size of the intermediate image may be adjusted to a preset size by padding.
The size of the intermediate image is compared with the preset size, and the size of the intermediate image larger than the preset size and the size of the intermediate image smaller than the preset size are respectively adjusted by different methods, so that the size of the region of interest in the obtained target image is kept unchanged, the phenomenon of image distortion is avoided, and the accuracy of the type recognition model in recognizing the target image is ensured.
Optionally, the method for adjusting the size of the intermediate image further includes: and inputting the intermediate image and the preset size into an image conversion model, and converting the intermediate image into a target image with the preset size through the image conversion model, wherein the optional image conversion model can be a neural network model.
S240, inputting the target image into the trained type recognition model, and determining the image type of the image of the region of interest.
And inputting the target image with the preset size and the preset shape into the trained type recognition model, so as to obtain the more accurate image type of the image of the region of interest.
Optionally, the training method of the type recognition model includes: acquiring a sample image comprising regions of interest and image types of the regions of interest of the sample image; extracting the outline of the region of interest of the sample image, generating a sample outline image, and generating a sample region image according to the sample outline image and the sample image; and converting the sample area image into a target sample image with a preset size and a preset shape, and performing iterative training on a type recognition model to be trained based on the target sample image and the image type corresponding to the target sample image to obtain the type recognition model based on the image type recognition function. The method comprises the steps of obtaining a sample image of an interested area, obtaining the image type of the interested area in each sample image, extracting the outline of the interested area in the sample image, setting the size of the sample image according to the outline size of the interested area, enabling the sample outline image to contain all the outlines of the interested area, setting the pixel value of an area inside the interested area to be 1 in the sample outline image, and setting the pixel value of an area outside the interested area to be 0. And performing image cutting operation on the sample image comprising the region of interest according to the sample outline image to obtain a sample cut image. Multiplying the sample cutting image and the sample contour image to obtain a sample area image, adjusting the size and the shape of the sample area image according to the method to enable the sample area image to be converted into a preset size and a preset shape to obtain a target sample image, inputting the target sample image into a type recognition model to be trained to obtain a prediction recognition result, comparing the prediction recognition result with an image type corresponding to the target sample image, calculating a loss function when the prediction recognition result is different from the image type corresponding to the target sample image, reversely inputting the loss function into the type recognition model to be trained, and adjusting network parameters in the type recognition model based on a gradient descent method. And iteratively executing the training method until the training of the preset times is finished, and determining that the training of the type recognition model is finished. And inputting the target image into a trained type recognition model, and determining the image type of the image of the region of interest.
According to the technical scheme of the embodiment of the invention, the image to be identified is obtained, and the outline of the region of interest in the image to be identified is extracted to generate the outline image; the image to be recognized is conveniently cut according to the contour image; obtaining an interested area image according to the outline image and the image to be identified; excluding the interference of other area images except the region of interest in the image to be identified on the type identification of the image; and converting the region-of-interest image into a target image with a preset size and a preset shape. The method has the advantages that the type recognition model can better classify the images of the region of interest, the image classification result is more accurate, the target image is input into the trained type recognition model, the image type of the image of the region of interest is determined, the problem that the efficiency of manually identifying the region of interest in the image to be recognized and further classifying the image to be recognized is low is solved, the type of the region of interest in the image is rapidly determined, and the medical image diagnosis efficiency is improved.
EXAMPLE III
Fig. 7 is a structural diagram of an image recognition apparatus according to a third embodiment of the present invention, where the image recognition apparatus includes: a contour image acquisition module 310, a region of interest image acquisition module 320, and an image type determination module 330.
The contour image acquiring module 310 is configured to acquire an image to be identified, extract a contour of a region of interest in the image to be identified, and generate a contour image; a region-of-interest image obtaining module 320, configured to obtain a region-of-interest image according to the contour image and the image to be identified; an image type determining module 330, configured to input the region of interest image into a trained type recognition model, and determine an image type of the region of interest image.
In the technical solution of the above embodiment, the contour image obtaining module 310 includes:
the contour extraction unit is used for extracting the contour of the interest region in the image to be identified;
and the contour image generating unit is used for generating a contour image according to the size of the contour, wherein in the contour image, the pixel point of the area inside the contour of the interested area is set to be 1, and the pixel point of the area outside the contour of the interested area is set to be 0.
In the technical solution of the above embodiment, the region-of-interest image obtaining module 320 includes:
the image cutting unit is used for carrying out cutting operation on the image to be identified according to the outline image to obtain a cut image, wherein the size of the cut image meets the required size of the image of the region of interest;
and the image multiplication unit is used for multiplying the cutting image and the contour image to obtain the interested area image.
In the technical solution of the above embodiment, the image recognition apparatus further includes:
the target image acquisition module is used for converting the region-of-interest image into a target image with a preset size and a preset shape;
accordingly, the image type determining module 330 includes:
and the target image input unit is used for inputting the target image into the trained type recognition model.
In the technical solution of the above embodiment, the target image obtaining module includes:
the pixel point filling unit is used for filling pixel points into the interested region image and converting the interested region image into an intermediate image in the preset shape;
and the size adjusting unit is used for adjusting the size of the intermediate image according to the size of the intermediate image and the preset size to obtain a target image.
In the technical solution of the above embodiment, the size adjusting unit is specifically configured to: determining an interpolation scaling of the intermediate image based on the size of the intermediate image and the preset size, and carrying out size adjustment on the intermediate image based on the interpolation scaling to obtain a target image; alternatively, the first and second electrodes may be,
if the size of the intermediate image is smaller than the preset size, performing pixel filling on the intermediate image based on the preset size, and if the size of the intermediate image is larger than the preset size, performing image contraction on the intermediate image based on the preset size to obtain a target image; alternatively, the first and second electrodes may be,
and inputting the intermediate image and the preset size into an image conversion model, and acquiring a target image which is output by the image conversion model and meets the preset size.
Optionally, the training method of the type recognition model includes:
acquiring a sample image comprising regions of interest and image types of the regions of interest of the sample image;
extracting the outline of the region of interest of the sample image, generating a sample outline image, and generating a sample region image according to the sample outline image and the sample image;
and converting the sample area image into a target sample image with a preset size and a preset shape, and performing iterative training on a type recognition model to be trained based on the target sample image and the image type corresponding to the target sample image to obtain the type recognition model based on the image type recognition function.
Optionally, the image to be identified is a medical image of a target object, the region of interest is a calcified region in the medical image, and the image type includes a calcification level of the calcified region.
According to the technical scheme of the embodiment of the invention, the image to be identified is obtained, and the outline of the region of interest in the image to be identified is extracted to generate the outline image; the image to be recognized is conveniently cut according to the contour image; obtaining an interested area image according to the outline image and the image to be identified; excluding the interference of other area images except the region of interest in the image to be identified on the type identification of the image; the image of the region of interest is input into a trained type recognition model, the image type of the image of the region of interest is determined, the problem that the efficiency is low when the region of interest in the image to be recognized is recognized manually and then the image to be recognized is classified is solved, the type of the region of interest in the image is determined rapidly, and the effect of improving the diagnosis efficiency of medical images is achieved.
The image recognition device provided by the embodiment of the invention can execute the image recognition method provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects of the execution method.
Example four
Fig. 8 is a schematic structural diagram of an image recognition apparatus according to a fourth embodiment of the present invention, as shown in fig. 8, the image recognition apparatus includes a processor 410, a memory 420, an input device 430, and an output device 440; the number of the processors 410 in the image recognition device may be one or more, and one processor 410 is taken as an example in fig. 8; the processor 410, the memory 420, the input device 430 and the output device 440 in the image recognition apparatus may be connected by a bus or other means, and fig. 8 illustrates the connection by a bus as an example.
The memory 420 serves as a computer-readable storage medium that may be used to store software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the image recognition apparatus method in the embodiments of the present invention (e.g., the contour image acquisition module 310, the region-of-interest image acquisition module 320, and the image type determination module 330 in the image recognition apparatus device). The processor 410 executes various functional applications and data processing of the image recognition device by executing software programs, instructions and modules stored in the memory 420, i.e., implements the image recognition device method described above.
The memory 420 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the terminal, and the like. Further, the memory 420 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, memory 420 may further include memory located remotely from processor 410, which may be connected to the image recognition device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input means 430 may be used to receive input numeric or character information and generate key signal inputs related to user settings and function control of the image recognition apparatus. The output device 440 may include a display device such as a display screen.
EXAMPLE five
An embodiment of the present invention further provides a storage medium containing computer-executable instructions, which when executed by a computer processor, perform an image recognition apparatus method, the method including:
acquiring an image to be identified, and extracting the outline of an interested area in the image to be identified to generate an outline image;
obtaining an interested area image according to the outline image and the image to be identified;
and inputting the image of the region of interest into a trained type recognition model, and determining the image type of the image of the region of interest.
Of course, the storage medium containing the computer-executable instructions provided by the embodiments of the present invention is not limited to the method operations described above, and may also perform related operations in the method for image recognition device provided by any embodiments of the present invention.
From the above description of the embodiments, it is obvious for those skilled in the art that the present invention can be implemented by software and necessary general hardware, and certainly, can also be implemented by hardware, but the former is a better embodiment in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute the methods according to the embodiments of the present invention.
It should be noted that, in the embodiment of the image recognition device apparatus, the included units and modules are only divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be realized; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (11)
1. An image recognition method, comprising:
acquiring an image to be identified, and extracting the outline of an interested area in the image to be identified to generate an outline image;
obtaining an interested area image according to the outline image and the image to be identified;
and inputting the image of the region of interest into a trained type recognition model, and determining the image type of the image of the region of interest.
2. The method according to claim 1, wherein the extracting the contour of the region of interest in the image to be identified generates a contour image, comprising:
extracting the outline of the region of interest in the image to be identified;
and generating a contour image according to the size of the contour, wherein in the contour image, the pixel points of the region inside the contour of the region of interest are set to be 1, and the pixel points of the region outside the contour of the region of interest are set to be 0.
3. The method according to claim 1, wherein the obtaining of the region-of-interest image from the contour image and the image to be identified comprises:
carrying out cutting operation on the image to be identified according to the outline image to obtain a cut image, wherein the size of the cut image meets the required size of the image of the region of interest;
and multiplying the cutting image and the contour image to obtain the region-of-interest image.
4. The method of claim 1, wherein prior to inputting the region of interest image into the trained type recognition model, the method further comprises:
converting the region-of-interest image into a target image with a preset size and a preset shape;
correspondingly, the inputting the region-of-interest image into the trained type recognition model includes:
and inputting the target image into a trained type recognition model.
5. The method according to claim 4, wherein the converting the region-of-interest image into a target image of a preset size and a preset shape comprises:
filling pixel points in the interested region image, and converting the interested region image into an intermediate image in the preset shape;
and adjusting the size of the intermediate image according to the size of the intermediate image and the preset size to obtain a target image.
6. The method according to claim 5, wherein the adjusting the size of the intermediate image according to the size of the intermediate image and the preset size to obtain the target image comprises:
determining an interpolation scaling of the intermediate image based on the size of the intermediate image and the preset size, and carrying out size adjustment on the intermediate image based on the interpolation scaling to obtain a target image; alternatively, the first and second electrodes may be,
if the size of the intermediate image is smaller than the preset size, performing pixel filling on the intermediate image based on the preset size, and if the size of the intermediate image is larger than the preset size, performing image contraction on the intermediate image based on the preset size to obtain a target image; alternatively, the first and second electrodes may be,
and inputting the intermediate image and the preset size into an image conversion model, and acquiring a target image which is output by the image conversion model and meets the preset size.
7. The method of claim 1, wherein the training method of the type recognition model comprises:
acquiring a sample image comprising regions of interest and image types of the regions of interest of the sample image;
extracting the outline of the region of interest of the sample image, generating a sample outline image, and generating a sample region image according to the sample outline image and the sample image;
and converting the sample area image into a target sample image with a preset size and a preset shape, and performing iterative training on a type recognition model to be trained based on the target sample image and the image type corresponding to the target sample image to obtain the type recognition model based on the image type recognition function.
8. The method according to any one of claims 1-7, wherein the image to be identified is a medical image of a target object, the region of interest is a calcified region in the medical image, and the image type includes a calcification level of the calcified region.
9. An image recognition apparatus, comprising:
the contour image acquisition module is used for acquiring an image to be identified and extracting the contour of an interested area in the image to be identified to generate a contour image;
the interesting region image acquisition module is used for obtaining an interesting region image according to the outline image and the image to be identified;
and the image type determining module is used for inputting the image of the region of interest into a trained type recognition model and determining the image type of the image of the region of interest.
10. An image recognition device, characterized in that the electronic device comprises:
one or more processors;
storage means for storing one or more programs;
when executed by the one or more processors, cause the one or more processors to implement the image recognition method of any of claims 1-8.
11. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the image recognition method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011501530.6A CN112541907A (en) | 2020-12-17 | 2020-12-17 | Image identification method, device, server and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011501530.6A CN112541907A (en) | 2020-12-17 | 2020-12-17 | Image identification method, device, server and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112541907A true CN112541907A (en) | 2021-03-23 |
Family
ID=75018983
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011501530.6A Pending CN112541907A (en) | 2020-12-17 | 2020-12-17 | Image identification method, device, server and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112541907A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115078430A (en) * | 2022-06-10 | 2022-09-20 | 水木未来(北京)科技有限公司 | Method and device for determining quality of support film of grid of cryoelectron microscope |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102005030646A1 (en) * | 2005-06-30 | 2007-01-11 | Siemens Ag | Method for contour visualization of regions of interest in 2D fluoroscopic images |
| CN104751147A (en) * | 2015-04-16 | 2015-07-01 | 成都汇智远景科技有限公司 | Image recognition method |
| CN109165646A (en) * | 2018-08-16 | 2019-01-08 | 北京七鑫易维信息技术有限公司 | The method and device of the area-of-interest of user in a kind of determining image |
| CN110148192A (en) * | 2019-04-18 | 2019-08-20 | 上海联影智能医疗科技有限公司 | Medical image imaging method, device, computer equipment and storage medium |
| CN110675464A (en) * | 2019-09-20 | 2020-01-10 | 和宇健康科技股份有限公司 | Medical image processing method and device, server and storage medium |
| CN110766659A (en) * | 2019-09-24 | 2020-02-07 | 西人马帝言(北京)科技有限公司 | Medical image recognition method, apparatus, device and medium |
-
2020
- 2020-12-17 CN CN202011501530.6A patent/CN112541907A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102005030646A1 (en) * | 2005-06-30 | 2007-01-11 | Siemens Ag | Method for contour visualization of regions of interest in 2D fluoroscopic images |
| CN104751147A (en) * | 2015-04-16 | 2015-07-01 | 成都汇智远景科技有限公司 | Image recognition method |
| CN109165646A (en) * | 2018-08-16 | 2019-01-08 | 北京七鑫易维信息技术有限公司 | The method and device of the area-of-interest of user in a kind of determining image |
| CN110148192A (en) * | 2019-04-18 | 2019-08-20 | 上海联影智能医疗科技有限公司 | Medical image imaging method, device, computer equipment and storage medium |
| CN110675464A (en) * | 2019-09-20 | 2020-01-10 | 和宇健康科技股份有限公司 | Medical image processing method and device, server and storage medium |
| CN110766659A (en) * | 2019-09-24 | 2020-02-07 | 西人马帝言(北京)科技有限公司 | Medical image recognition method, apparatus, device and medium |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115078430A (en) * | 2022-06-10 | 2022-09-20 | 水木未来(北京)科技有限公司 | Method and device for determining quality of support film of grid of cryoelectron microscope |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP2378978B1 (en) | Method and system for automated generation of surface models in medical images | |
| CN107480677B (en) | Method and device for identifying interest region in three-dimensional CT image | |
| CN108898595B (en) | Construction method and application of positioning model of focus region in chest image | |
| CN109493325B (en) | Tumor heterogeneity analysis system based on CT images | |
| US6795521B2 (en) | Computer-aided diagnosis system for thoracic computer tomography images | |
| CN107527341B (en) | Method and system for processing angiography image | |
| Chakraborty et al. | Automatic detection of pectoral muscle using average gradient and shape based feature | |
| CN108062749B (en) | Identification method and device for levator ani fissure hole and electronic equipment | |
| KR102206621B1 (en) | Programs and applications for sarcopenia analysis using deep learning algorithms | |
| JP2015066311A (en) | Image processor, image processing method, program for controlling image processor, and recording medium | |
| CN110880177A (en) | Image identification method and device | |
| Kaur et al. | Computer-aided diagnosis of renal lesions in CT images: a comprehensive survey and future prospects | |
| Moon et al. | Quantitative analysis for breast density estimation in low dose chest CT scans | |
| CN114332132A (en) | Image segmentation method and device and computer equipment | |
| CN112541907A (en) | Image identification method, device, server and medium | |
| Cerneaz et al. | Finding curvilinear structures in mammograms | |
| Mustafa et al. | Mammography image segmentation: Chan-Vese active contour and localised active contour approach | |
| CN111223090A (en) | Identification system of tumor image in human lung CT image | |
| Joda et al. | Digital mammogram enhancement based on automatic histogram clipping | |
| CN111383323B (en) | Ultrasonic imaging method and system and ultrasonic image processing method and system | |
| Gayathri et al. | An automatic identification of lung cancer from different types of medical images | |
| Kumar et al. | Pulmonary nodules diagnosis from x-ray imaging using image processing | |
| Abdolali et al. | Fully automated detection of the mandibular canal in cone beam CT images using Lie group based statistical shape models | |
| JP2022133479A (en) | Prognosis estimation device and prognosis estimation method | |
| Feudjio et al. | Automatic Extraction of breast region in raw mammograms using a combined strategy |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210323 |