Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples are not representative of all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with some aspects of the present disclosure as detailed in the accompanying claims.
Embodiments of the present disclosure are described in detail below with reference to the attached drawing figures.
The present disclosure provides a new scanning method applied to a medical system, which may include a CT apparatus, an MRI apparatus, a PET/CT apparatus, and the like. By executing the method, the medical system can automatically position the body part in the positioning sheet and automatically set the scanning visual field.
Embodiments of the present disclosure are described in detail below with reference to the attached drawing figures.
Fig. 1 is a flow chart of a scanning method shown in an exemplary embodiment of the present disclosure, which may include the steps of:
in step 101, a set body part is identified from a slice taken for an object to be scanned.
The somatic part refers to organs or bones in the body, such as the lung, liver, stomach, ilium, pubis, etc. The scanning system has a function of automatically identifying the body part from the splines.
The body part is set as the body part for determining the scanning visual field under the specific scanning requirement.
In clinical scans, the body can be divided into: the upper abdomen, middle abdomen and lower abdomen, the collection of the upper and middle abdomen is called the abdomen, and the lower abdomen is also called the pelvis. Anatomically, the abdomen goes from the diaphragm at the chest bottom to the anterior iliac spine; the pelvis is formed by the sacrum, coccyx, hip bone and pubic bone.
Based on different scanning requirements, the scanning can be classified into abdominal scanning, pelvic scanning, abdominal-basin combined scanning, and the like, wherein the abdominal scanning is scanning performed for the abdomen, the pelvic scanning is scanning performed for the lower abdomen, and the abdominal-basin combined scanning is scanning performed for the abdomen and pelvis.
In performing an abdominal scan, the lower edge of the lung is used to define the upper edge of the abdomen, the upper edge of the ilium is used to define the lower edge of the abdomen, and the lung and ilium are used as the set body site. In performing a pelvic scan, the upper edge of the ilium is used to determine the upper edge of the pelvis, the lower edge of the pubis is used to determine the lower edge of the pelvis, and the ilium and pubis used as the set torso region.
In step 102, the position of the target body area associated with the set body part is determined based on the position of the set body part in the positioning sheet.
The target body region is used to determine a target scan field of view. For example, the target somatic region is an abdominal region, a pelvic region, an abdomen-basin joint region, or the like.
In one embodiment, the position of a boundary of the target body area may be determined based on the position of one of the set body parts in the splines, and the position of the other edge of the target body area may be determined based on the position of the other set body part in the splines, thereby determining the position of the target body area in the splines.
For example, in performing an abdominal scan, the location of the abdominal region may be determined based on the location of the lower edge of the lung and the location of the upper edge of the ilium in the splines. In performing a pelvic scan, the location of the pelvic region may be determined from the location of the upper edge of the ilium and the location of the lower edge of the pubic bone in the splines.
In one embodiment, the scanning system includes a trained body part recognition model for recognizing a body part from the splines, and in particular, the body part recognition model is for recognizing a corresponding set body part from the splines under a specified scanning requirement.
Based on this, the scanning system can recognize the set body part from the positioning sheet using the trained body part recognition model.
In the process of using the model, the positioning sheet can be input into the trained body part recognition model, a binary image output by the body part recognition model is obtained, the binary image comprises a first numerical value and a second numerical value, and a communication area formed by a plurality of first numerical values is determined as an area where the set body part is located. The connected region formed by the plurality of first values may be referred to as a foreground region and the connected region formed by the plurality of second values may be referred to as a background region.
After the binary image output by the part identification model is obtained, the scanning system determines a communication area formed by the first numerical value, and when the communication area corresponding to the non-body part, namely, the small communication area with interference exists, the small communication area can be deleted so as to remove the interference.
In this embodiment, the set body part is identified from the positioning sheet by using the model, and an accurate part identification result can be obtained based on the characteristics of the model.
In one embodiment, the scanning system determines a position of a set body site in the splines, and determines a position of a target body area associated with the set body site based on the position of the set body site in the splines.
There are various ways of determining the position of the set body part in the spacer, for example, as follows:
a first determination mode: when the structure of the set body part in the locating piece is complete, the position of the target body area can be determined according to the position of the edge of the set body part in the locating piece.
For example, in the splines shown in fig. 2, the pelvic structure is clear and complete, the position of the upper boundary of the pelvic region is determined based on the position of the upper edge of the ilium, such as the highest point of the upper edge of the ilium, a horizontal line is made across the highest point, the horizontal line is taken as the upper boundary of the pelvic region, and the position of the lower boundary of the pelvic region is determined based on the position of the lower edge of the pubis, such as the lowest point of the lower edge of the pubis, a horizontal line is made across the lowest point, and the horizontal line is taken as the lower boundary of the pelvic region. The widths of the upper and lower boundaries may be determined according to the width of the human body.
As another example, the splines shown in fig. 3 are very clear, with the location of the upper boundary of the abdominal region being determined based on the location of the lower edge of the lungs, and the location of the lower boundary of the abdominal region being determined based on the location of the upper edge of the ilium.
The second determination mode: when the set body part in the locating piece is incomplete, determining the estimated edge and the position of the estimated edge of the set body part in the locating piece according to the structure and the position of the set body part in the locating piece.
When the splines comprise only a partial structure of the set body part, the second determination may be used to determine the position of the set body part in the splines.
The body part is divided into a plurality of sub-parts, the structures and the positions of the different sub-parts have corresponding relations based on the physiological structural characteristics of the body part, and the structures and the positions of the other sub-parts can be estimated according to the structures and the positions of one sub-part. Based on this, the edge of the set body part (i.e., the estimated edge) and the position of the edge can be estimated from the partial structure of the set body part.
The predicted edges may include all edges of the set body part, or the predicted edges may include only target edges of the set body part, the target edges being used to determine the target body area.
For example, the locating piece shown in fig. 4 only includes a part of the structure of the lung, and the whole structure of the lung can be estimated according to the part of the structure of the lung, so as to obtain the estimated edge of the lung.
Third determination mode: when the set body part in the locating piece is incomplete, determining the estimated structure and the position of the estimated structure of the set body part in the locating piece according to the structures and the positions of other body parts in the locating piece.
When the splines comprise only a partial structure of the set body part or the splines do not comprise the set body part, a third determination may be made as to the position of the set body part in the splines.
For the body, the structures and positions of different body parts have a certain relationship, and the structure and the position of one body part can be estimated according to the structure and the position of the other body part. The sizes of the different body parts have a certain relation, and the size of one body part can be determined according to the size of the other body part and the pre-obtained size relation of the different body parts.
For example, when performing a joint abdominal-pelvic scan, the spacer shown in fig. 5 is obtained, and the presence of a loss of pubic structure in the spacer shown in fig. 5 may be used by the scanning system to predict pubic structure and position based on the structure and position of other bones in the pelvis, such as ilium.
Fourth determination mode: when the set body part in the locating piece is incomplete, determining the estimated size of the set body part according to the sizes of other body parts in the locating piece, determining the position of the other edge of the set body part in the locating piece according to the position and the estimated size of one edge of the set body part in the locating piece, and determining the position of the other edge of the set body part for determining the position of the target body area.
In a fourth mode of determination, the splines comprise only one edge of the set body part.
The sizes of different body parts have a certain relation, and the size (i.e. the estimated size) of the set body part can be determined according to the sizes of other body parts. Setting the estimated size of the body part may include setting the size of the body part in the height, width, etc. directions in the positioning sheet.
For example, the lower pubic portion of the splines shown in fig. 5 may be missing, the scanning system may predict the height of the pubic portion of the splines based on the size of other body parts, and may determine the location of the lower edge of the pubic portion of the splines based on the upper edge of the pubic portion of the splines and the predicted height of the pubic portion, the location of the lower edge of the pubic portion of the splines may be used to determine the lower boundary of the pelvic or abdominal region.
In step 103, a target scan field of view including the target body region is determined based on the location of the target body region.
The target scan field of view may coincide with the target body region or the target scan field of view may be larger than and include the target body region.
In step 104, a target body region of the subject to be scanned is scanned using the target scan field of view.
In one embodiment, the number of target body regions associated with the set body part is more than two, each corresponding to a target scan field of view. The scanning of the target body area of the object to be scanned using the target scanning field of view may be achieved by: and in response to outputting the position information of the target scanning vision field corresponding to each body area, and receiving the target position information determined from more than two pieces of position information, scanning the corresponding target body area of the object to be scanned by using the target scanning vision field corresponding to the target position information.
There are various ways of outputting the position information of the target scanning field corresponding to each body region, for example, displaying the frame of the target scanning field on the positioning sheet, or displaying the coordinates of the target scanning field in the positioning sheet on the positioning sheet.
The doctor can select the required scanning visual field to scan according to the requirement, so that the manual intervention on the scanning visual field is realized.
On the basis, the scanning system can also determine the use recommendation information of the target scanning vision fields corresponding to the target body areas according to the structural integrity of the target body areas in the positioning sheet, and output the use recommendation information of the target scanning vision fields for a doctor to select the target scanning vision fields for reference.
Generally, the higher the structural integrity of the target body region, the greater the recommended strength of use of the target body region. There are various forms of using the recommendation information, for example, using the recommendation level, the higher the recommendation level, the more recommended, and for example, using the recommendation score, the higher the recommendation score, the more recommended.
The embodiment of the disclosure provides a scanning method, which is characterized in that a set body part is identified from a locating plate shot for an object to be scanned, the position of the set body part in a root locating plate is determined, the position of a target body area related to the set body part is determined, a target scanning visual field comprising the target body area is determined according to the position of the target body area, and the target body area of the object to be scanned is scanned by using the target scanning visual field, so that the body part in the locating plate is automatically positioned by a scanning system, the body area is automatically positioned, the scanning visual field is automatically set, the operation of manually setting the scanning visual field by a doctor in the related technology is omitted, the clinical examination time is shortened, the clinical examination efficiency is improved, and the workload of the doctor is lightened.
Meanwhile, the scanning system sets the FOV by using a unified method, so that the acquired images tend to be standardized, the image quality difference caused by different use habits and operation methods is reduced, and the combined diagnosis and treatment and medical case communication is facilitated.
Because the scanning system automatically sets the FOV, the operation of manually setting the FOV by a doctor is omitted, the workload of the doctor is reduced, and meanwhile, the occurrence of incorrect setting of the FOV by the doctor is avoided.
Fig. 6 is a flowchart illustrating a method of training a body part recognition model according to an exemplary embodiment of the present disclosure, and referring to fig. 6, a scanning system may train the body part recognition model by:
in step 201, a sample image is acquired, the sample image having a tag indicating position information of a set body part in the sample image.
For example, the position information indicated by the tag may include contour position information of the set body part in the positioning sheet, and may be expressed in the form of coordinates.
The physician may manually label the splines to obtain a sample image with labels. For example, the physician may encircle the body part in the splines and label the name of the body part.
In the case of insufficient samples, to increase the robustness of the model, one or more of the following data enhancement operations may be performed on the tile samples: data enhancement operations such as overturning, translation, rotation, noise disturbance adding, color dithering and the like enrich the number of samples.
Generally, the body part recognition model has a function of recognizing multiple body parts, and in order to ensure that the body part recognition model can accurately recognize multiple body parts, the number of tags indicating various body part information in a sample set needs to be controlled, and the number of each tag cannot be reduced, otherwise, the body part recognition model has poor capability of recognizing the corresponding body part.
Corresponding to some body parts with smaller sizes, the body part identification model is easy to misidentify the background structure as the body part. In order to solve the above problems, the positioning sheet sample can be cut, the small-size body part is reserved, other image areas are cut off, so that the background structure is reduced, and the accuracy of identifying the small-size body part by using the cut sample image training model can be improved.
For example, the abdomen positioning slice sample comprises lung, ilium and pubis, the pubis size is small, the body part recognition model is easy to misidentify the background structure as the pubis structure, in order to reduce misidentification, a doctor cuts out the sample image of the abdomen, such as cutting out a lung image or cutting out a lung image and an ilium image, the cut sample image is input into the body part recognition model, and the body part recognition model has better pubis recognition capability through model parameter optimization.
In step 202, a sample image is input into a body part recognition model to be trained, and position information of a set body part output by the body part recognition model is obtained.
In step 203, parameters in the body part recognition model are adjusted based on the position information output from the body part recognition model and the position information indicated by the tag.
For example, the parameters in the body part identification model may be adjusted until the difference between the position information output by the body part identification model and the position information indicated by the tag is smaller than a preset difference, or until the number of times of adjustment of the parameters in the body part identification model reaches a preset number of times, the operation of adjusting the parameters may be stopped.
The method provided by the embodiment can be used for training the model, and after the training is finished, the body part identification model capable of identifying the set body part from the locating plate is obtained.
In one embodiment, for step 201, a sample image may be acquired by: firstly, acquiring an initial sample image, wherein the initial sample image is provided with a label, and the label indicates position information of a set body part in the initial sample image; next, noise is added to the initial sample image, and a sample image is obtained.
The body part recognition model is trained by using the sample image added with noise, so that the trained body part recognition model recognizes the set body part from the positioning sheet with low resolution and low quality, and has better body part recognition capability.
For example, in the spacer shown in fig. 7, the abdomen is blocked, the abdomen structure is unclear, the quality of the spacer is poor, and the upper edge of ilium can be accurately identified from the spacer with poor quality by using the body part identification model obtained by model training using the sample image to which noise is added.
For another example, in the spacer shown in fig. 4, the pelvis structure is blocked, the pelvis structure is unclear, and the image resolution is low, and for the body part recognition model obtained after model training using the sample image added with noise, the lower edge of the pubis can be accurately recognized from the low-resolution spacer by using the body part recognition model.
For another example, in the spacer shown in fig. 8, the ilium structure is unclear, and the upper edge of the ilium can be accurately identified from the low-resolution spacer by using the body part identification model obtained by performing model training using the sample image to which noise is added. The scanning system determines the scanning field of view used in scanning the abdominal-pelvic union region from the lower edge of the lung and the lower edge of the pubic bone, represented by the solid border, while the scanning system determines the scanning field of view used in scanning the pelvic region from the clear lower edge of the pubic bone and the identified upper edge of the ilium, represented by the dashed border. The doctor determines that two scan fields exist by looking at the two borders in the spacer.
In one embodiment, if the acquired sample image has large noise, the noise is not required to be added to the sample image, the body part identification model is trained by using the sample image with large noise, and the trained body part identification model can identify the set body part from the positioning sheet with low resolution and low quality and has better body part identification capability.
Corresponding to the foregoing scanning method, the present disclosure also provides embodiments of a scanning apparatus and an image processing device.
Referring to fig. 9, which is a schematic diagram of a scanning apparatus according to an exemplary embodiment of the present disclosure, the scanning apparatus is applied to a scanning system, and the scanning apparatus includes: a part identification module 31, a position determination module 32, a field of view determination module 33, and a body scanning module 34; wherein,,
the part identifying module 31 is configured to identify a set body part from the positioning sheet shot for the object to be scanned;
the position determining module 32 is configured to determine a position of a target body area associated with the set body part according to the position of the set body part in the positioning sheet;
the field of view determination module 33 is configured to determine a target scan field of view including the target body area based on the position of the target body area;
the body scanning module 34 is configured to scan the target body region of the object to be scanned using the target scan field of view.
Referring to fig. 10, which is a schematic diagram of a scanning system shown in an exemplary embodiment of the present disclosure, the system may include: a memory 320, a processor 330, and an external interface 340 connected by an internal bus 310.
Wherein, the external interface 340 is used for acquiring data;
a memory 320 for storing machine-readable instructions corresponding to the scan;
a processor 330 for reading the machine readable instructions on the memory 320 and executing the instructions to perform the following operations:
identifying a set body part from a positioning sheet shot for an object to be scanned;
determining a position of a target body area associated with the set body part according to the position of the set body part in the positioning sheet;
determining a target scan field of view including the target body region based on the location of the target body region;
the target body region of the object to be scanned is scanned using the target scan field of view.
In the disclosed embodiments, the computer-readable storage medium may take many forms, such as, in different examples, the machine-readable storage medium may be: RAM (Radom Access Memory, random access memory), volatile memory, non-volatile memory, flash memory, a storage drive (e.g., hard drive), a solid state drive, any type of storage disk (e.g., optical disk, dvd, etc.), or a similar storage medium, or a combination thereof. In particular, the computer readable medium may also be paper or other suitable medium capable of printing a program. Using these media, the programs may be electronically captured (e.g., optically scanned), compiled, interpreted, and otherwise processed in a suitable manner, and then stored in a computer medium.
The foregoing description of the preferred embodiments of the present disclosure is not intended to limit the disclosure, but rather to cover all modifications, equivalents, improvements and alternatives falling within the spirit and principles of the present disclosure.