KR101982941B1 - Method and Apparatus for removing artifact in CT image using fuzzy neural network - Google Patents

Abstract

The present invention relates to a method and an apparatus for removing an artifact from a CT image, and more particularly, to a method and an apparatus for removing an artifact from a CT image using a fuzzy neural network, which acquires raw data obtained by tomography of an object from a photon counting detector, detects artifact data forming a linear shape by converting the raw data including an artifact into polar coordinates, forms output data that converges to a predetermined value by inputting the artifact data to a fuzzy neural network and applying a fuzzy membership function thereto, and obtains correction data by converting the output data into cartesian coordinates, thereby detecting a more accurate artifact by converting an image acquired from the photon counting detector, and providing an image from which the detected artifact is accurately and conveniently removed by using the learning ability of the fuzzy neural network.

Description

Method and apparatus for removing virtual image of CT image using fuzzy neural network {Method and Apparatus for removing artifact in CT image using fuzzy neural network}

The present invention relates to a method and apparatus for removing a virtual image from a CT image, and more particularly, to a method and apparatus for removing an artifact of a CT image using a fuzzy neural network that removes a circular virtual image appearing on a CT image using a learning ability of a fuzzy neural network. It is about.

Diagnostic medical radiographic imaging devices that provide digital imaging are advancing not only in technical aspects but also in clinical applications. Needless to say, many studies are being conducted to develop new diagnostic methods, such as PET-CT and PET-MR, which combine existing diagnostic methods.

Among CT medical imaging devices, CT scans (hereinafter referred to as 'CT') use MRP (Multi Planar Reconstruction), similar to MR or PET, such as transverse plane, sagittal plane and coronal plane. In addition to providing two-dimensional tomographic and three-dimensional images of various reference planes, such as (coronal plane), the contrast resolution is good, and the scan time is short, so it is a very important area in the current medical radiographic diagnosis field. .

CT scans the anatomical tomography image inside the human body by reconstructing the spatial distribution of material properties between the X-ray source and the detector using mathematical methods of data obtained from projection images taken from several different angles. to provide.

Image quality is a very important factor for accurate disease diagnosis, and image quality in CT is defined as how clear and accurate images of the human microstructure can be visualized and affect the quality of CT images. Factors affecting the accuracy of the CT values are noise, noise, spatial and contrast resolution, artifacts, dose, and device accuracy.

Among these factors, artifacts are obstacle shadows that deteriorate the quality of images and reduce the ability to observe microscopic parts by disturbing or obstructing the correct CT image due to statistical errors that occur regardless of the inspection purpose. The projection image itself, which is the input data used for image reconstruction, is defined by differences in gain between each detector array element, patient movement, scattering lines caused by metals, X-ray beam hardening and partial volume effects while penetrating the patient. Because the projection information and other errors caused by the actual anatomical structure of the patient generated due to the error is included in the reconstructed CT image will have to appear artificial (artifact). In addition, a virtual method such as loss of information of a subject or blurring of an edge of a reconstructed image is generated by a mathematical method of a back projection process.

In order to remove such a virtual image, a low pass filter, a high pass filter, and the like are used, and various types of filters are used for each, and filters of a combination of them are also used. However, there is a problem that the characteristics of each filter is different, and the use of the filter on the CT image is very large.

Also, before reconstructing the 3D image to remove the virtual image, the position of the vertical stripe corresponding to the virtual image is detected in the sinogram image, and then the position value of the virtual image detected by the interpolation method is replaced. Although many detectors are used to calibrate the pixels themselves, the process of finding and correcting various virtual images is not easy, and it replaces the original values with the interpolation of surrounding pixel values. There is a problem that information loss occurs.

On the other hand, fuzzy logic is a useful means for expressing information like human language, and neural network is used effectively in various control fields by modeling the human brain and storing and using information by learning. Is used to accurately predict the result under any conditions and conditions based on the learned results. It is an adaptive decision support tool that combines neural network and fuzzy set theory for pattern classification, diagnosis, and prediction. Fuzzy Neural Network (FNN), a support tool, has been proposed.

Recently, many studies have been conducted to use fuzzy neural networks in various fields such as removing noise from an image or detecting / recognizing certain feature objects.

1. Republic of Korea Patent No. 10-1717433 'Article correction method by beam hardening phenomenon in X-ray computed tomography environment'

The present invention has been made to solve the above problems, by converting the image obtained from the photon counting detector to detect a more accurate virtual image and can provide an image that is accurately and conveniently removed the detected virtual image using the learning ability of the fuzzy neural network An object of the present invention is to provide a method and apparatus for removing artifacts from a CT image using a fuzzy neural network.

In accordance with another aspect of the present invention, there is provided a method of removing a virtual image of a CT image using a fuzzy neural network, the method including: obtaining raw data including an artifact generated by tomography of an object from a photon counting detector; Detecting the virtual data having a line shape by converting the raw data by polar coordination; Emphasizing the virtual image forming a line by gray level transformation of the virtual image data; Inputting grayscale-converted virtual image data into a fuzzy neural network and applying fuzzy membership function to form output data that converges to a predetermined value; And converting the output data into Cartesian coordinates to obtain correction data.

delete

delete

delete

delete

delete

The above-described method for removing virtual images of CT images using the fuzzy neural network of the present invention is composed of a computer-readable recording medium having recorded thereon a program for executing on a computer. The recording medium is at least one of a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device.

On the other hand, the virtual image removal apparatus of the CT image using the fuzzy neural network of the present invention for achieving the above object to obtain the raw data including the artifact generated by tomography of the object from the photon counting detector (photon counting detector) Image acquisition unit; An image converter configured to detect the virtual data having a linear shape by converting the raw data into polar coordinates, and converting the output data formed using the virtual data into Cartesian coordinates to obtain correction data; An image compensator for emphasizing the virtual image forming a line by gray level transformation of the virtual image data; And an image processor configured to input grayscale-converted virtual image data into a fuzzy neural network and apply fuzzy membership function to form output data that converges to a predetermined value.

delete

delete

delete

delete

delete

According to the present invention, it is possible to detect a more accurate virtual image by converting the image obtained from the photon count detector and to remove the detected virtual image accurately and conveniently through the fuzzy neural network, thereby improving the quality of the image provided to the user The diagnostic efficiency can be improved.

In addition, it can be used to make accurate diagnosis on various parts of the chest, as well as knee joints, back bones of the hands, sinuses, breasts, etc., and the data accumulated through fuzzy neural networks can be used in various medical fields. will be.

1 is a view showing an apparatus for removing artifacts of CT images using a fuzzy neural network according to the present invention,
2 is a view showing a state in which polar data is converted to raw data applied to the present invention;
3 is a view showing the structure of a fuzzy neural network applied to the present invention;
4 is a view showing an example of a pixel connectivity operation applied to the present invention;
FIG. 5 is a view illustrating a state in which gray scale conversion and pixel connectivity are performed on virtual image data applied to the present invention; FIG.
6 is a diagram illustrating a state in which hybrid interpolation is performed on virtual image data applied to the present invention;
7 to 9 are views showing an embodiment of a method for removing artifacts of CT images using the fuzzy neural network according to the present invention.

Advantages and features of the present invention, and methods of achieving them will be apparent with reference to the embodiments described below in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but can be implemented in various different forms, and only the embodiments make the disclosure of the present invention complete, and the general knowledge in the art to which the present invention belongs. It is provided to fully inform the person having the scope of the invention, which is defined only by the scope of the claims. Throughout the specification, like reference numerals refer to like elements.

Terms used herein will be briefly described and the present invention will be described in detail.

The terms used in the present invention have been selected as widely used general terms as possible in consideration of the functions in the present invention, but this may vary according to the intention or precedent of the person skilled in the art, the emergence of new technologies and the like. In addition, in certain cases, there is also a term arbitrarily selected by the applicant, in which case the meaning will be described in detail in the description of the invention. Therefore, the terms used in the present invention should be defined based on the meanings of the terms and the contents throughout the present invention, rather than the names of the simple terms.

When any part of the specification is to "include" any component, this means that it may further include other components, except to exclude other components unless otherwise stated. In addition, the term "part" as used herein refers to a hardware component, such as software, FPGA or ASIC, and "part" plays certain roles. But wealth is not limited to software or hardware. The 'unit' may be configured to be in an addressable storage medium, or may be configured to play one or more processors. Thus, as an example, a "part" refers to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, procedures, Subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays and variables. The functionality provided within the components and "parts" may be combined into a smaller number of components and "parts" or further separated into additional components and "parts".

In the present specification, when one component "transmits" or "provides" data or a signal to another component, the component may directly transmit the data or signal to another component, and at least one another This means that a component can transmit data or a signal to another component.

Hereinafter, with reference to the accompanying drawings will be described in detail to be easily carried out by those of ordinary skill in the art to which the present invention pertains. As those skilled in the art would realize, the described embodiments may be modified in various different ways, all without departing from the spirit or scope of the present invention. In the drawings, parts irrelevant to the description are omitted in order to clearly describe the present invention.

As used herein, "image" may mean multi-dimensional data composed of discrete image elements (eg, pixels in a 2D image and voxels in a 3D image). For example, the image may include a medical image of the object obtained by the CT imaging apparatus.

In the present specification, a “CT (Computed Tomography) image” may mean a composite image of a plurality of X-ray images obtained by rotating about at least one axis of an object and photographing the object.

As used herein, an "object" may be a person or an animal, or part or all of a person or an animal. For example, the subject may include at least one of organs such as the liver, heart, uterus, brain, breast, abdomen, and blood vessels. In addition, the "object" may be a phantom. Phantom means a material having a volume very close to the density and effective atomic number of an organism, and may include a sphere phantom having properties similar to the body.

As used herein, a "user" may be a doctor, a nurse, a clinical pathologist, a medical imaging professional, or the like, and may be a technician who repairs a medical device, but is not limited thereto.

1 is a view showing an apparatus for removing an artifact of a CT image using a fuzzy neural network according to the present invention, the apparatus for removing an artifact of a CT image using a fuzzy neural network includes a photon count detector 100, a virtual image removal device 200, The virtual image removing apparatus 200 may include an image obtaining unit 210, an image converting unit 220, and an image processing unit 240, and may further include an image compensating unit 230.

The photon counting detector 100 is for receiving a CT image obtained by a user tomography of an object. The photon counting detector 100 detects an X-ray transmitted through the object, and detects an X-ray transmitted through the object. After generating a signal, the digital signal is generated to generate raw data obtained by tomography of the object. The photon count detector 100 transmits the raw data to the image acquisition unit 210 through a communication network.

The communication network is a high-speed network of a large communication network capable of large-capacity, long-distance voice and data services, and may be a next-generation wired and wireless network for providing an Internet or high-speed multimedia service. When the wired / wireless communication network is a mobile communication network, it may be a synchronous mobile communication network or an asynchronous mobile communication network. An example of an asynchronous mobile communication network is a communication network of a wideband code division multiple access (WCDMA) scheme. In this case, although not shown in the figure, the wired or wireless communication network may include a Radio Network Controller (RNC). Meanwhile, although the WCDMA network is taken as an example, it may be an IP network based on a next generation communication network such as 3G LTE network, 4G network, or other IP. The communication network plays a role of mutually transferring signals and data between a database, an authentication server, and other systems included in the present invention, including the photon counting detector 100 and the virtual image removing apparatus 200.

The image acquisition unit 210 receives raw data obtained by tomography imaging an object from the photon counting detector 100 and transmits the received raw data to the image converter 220. In this case, the image acquisition unit 210 classifies and detects X-rays from the photon count detector 100 for each of a plurality of energy bands, obtains raw data for the object for each of the plurality of energy bands, and selects the corresponding low data. The image converter 220 may transmit the image.

For example, the photon counting detector 100 may detect an X-ray by dividing the first band of 10 keV to 22 keV, the second band of 22 keV to 36 keV, and the third band of 36 keV to 50 keV, and acquire the image. The unit 210 may obtain raw data for an object classified into a first band of 10 keV to 22 keV, a second band of 22 keV to 36 keV, and a third band of 36 keV to 50 keV.

FIG. 2 is a view illustrating a state in which polar data is converted to polar data applied to the present invention, and the image converter 220 polarizes raw data including an artifact as shown in FIG. 2. coordination) It detects virtual data which is transformed and forms a line.

At this time, the morphological edge detection operator can detect the virtual data in the form of a line in the raw data converted to polar coordinates by using Equation 1 below, where B in Equation 1 represents a structural element, A Means the image to be applied.

In addition, the image converter 220 obtains correction data by converting Cartesian coordinates of output data formed by the image processor 240 to be described later. That is, after performing the operation of removing the virtual image by polar coordinate transformation of the raw data of the tomography image of the object, the image is converted to Cartesian coordinates again to improve the quality of the image provided to the user whose virtual image is removed through the correction data. Efficiency can be increased.

To this end, the image processing unit 240 in the virtual image removal apparatus of the CT image using the fuzzy neural network of the present invention receives the virtual data detected by the image conversion unit 220 to purify the fuzzy neural network (fuzzy neural network) The input image is input to the fuzzy membership function and the output data converges to a certain value to accurately and conveniently remove the detected virtual image.

FIG. 3 is a diagram illustrating a structure of a fuzzy neural network according to the present invention, wherein the fuzzy neural network is an input measurement, a fuzzy partition layer, and a front combination layer as shown in FIG. 3. layer, inference layer, and output layer.

The input layer is composed of n input nodes and receives n virtual data from each node. In addition, the virtual image data input by the input layer may be virtual image data obtained by converting only polar coordinates by the image converting unit 220, but the virtual image data is corrected by the image correction unit 230 to be described later. Can be.

The fuzzy partition layer applies an input function to n virtual data to form an input fuzzy set composed of a plurality of fuzzy data. For example, as shown in FIG. 3, one virtual data forms five input fuzzy sets. can do.

The full addition layer classifies the fuzzy data by assigning the input fuzzy set formed by the at least two virtual image data according to the distance to the specific fuzzy data. Accordingly, as shown in FIG. 3, when two virtual data forms more than n m fuzzy sets, the fuzzy data forms more than m p pieces.

The inference layer selects the inference data by applying a fuzzy rule base to the classified fuzzy data to form an output fuzzy set. The fuzzy rulebase means a formula and a rule of fuzzy theory, and the output fuzzy set forms less than q fuzzy data having p.

The output layer forms output data through unpurging the output fuzzy set, and the output data converges to a predetermined value.

Meanwhile, as described above, the apparatus for removing a virtual image of a CT image using the fuzzy neural network of the present invention may further include an image corrector 230 capable of correcting the virtual image data.

The image correction unit 230 may perform a function of emphasizing the virtual image forming a line by gray level transformation of the virtual image data. That is, by changing the concentration of each point so that the concentration distribution of the virtual image or the concentration distribution of the portions other than the virtual image is changed to a specific concentration distribution, the virtual image forming a line may be raised.

FIG. 4 is a diagram illustrating an example of a pixel connectivity operation applied to the present invention, and the image compensator 230 extracts a feature point of the virtual image data as shown in FIG. 4 and grasps a connection form of the feature point. The pixel connectivity function of selecting and removing pseudo-feature points may be performed. Here, the pseudo feature point means a feature extracted as a feature point but not a feature point in an actual image.

Referring to an example of a process of removing pseudo-feature points with reference to FIG. 4, feature points (whole 1, whole 2, whole 3) are extracted from the virtual image, and 4 connectivity is applied to detect feature points. The feature points of whole 1 and whole 3 excepted are removed as pseudo feature points. Then, by applying 2 connectivity, the feature points of whole 2 are removed by pseudo feature points to form an extracted line for whole 1, which is a pixel having connectivity.

FIG. 5 is a view illustrating a state in which gray level conversion and pixel connectivity are performed on virtual image data applied to the present invention, and the image correction unit 230 performs both gray level conversion and pixel connectivity. Can produce the same result as 5.

Also, the image interpolation 230 may perform interpolation, which is a method of estimating an unknown value using known data values. The image interpolation 230 may perform various types of interpolation methods. For example, high-order interpolation (bicubic-interpolation) that selects a target pixel from virtual data and takes an average value of other pixels surrounding the target pixel in the target pixel. ), And a hybrid interpolation method that combines polynomial fitting-interpolation to obtain a function value for an arbitrary position using an approximation function composed of polynomials.

Higher-order interpolation, which is performed when the image and / or image is changed in size, calculates and processes the extra space created by magnification during enlargement, for example. Simple but good video expansion. Higher-order interpolation also takes the average of the data (RGB or YCbCr) values of the colors of the 15 (4X4) pixels surrounding the pixels to be displayed, and changes the values of other pixels that are the reference of the target pixel to be displayed. However, since the change is very small, it is possible to minimize the loss of information on the original image.

In the case of polynomial approximation interpolation, the operation may be performed using a first order, a second order, a third order polynomial, and the like. In the present invention, since the higher order interpolation is performed together, the interpolation may be complemented. FIG. 6 is a diagram illustrating a state in which hybrid interpolation is performed on virtual image data applied to the present invention, and thus corrected virtual image data can be obtained.

7 to 9 are views illustrating an embodiment of a method for removing an artifact of a CT image using a fuzzy neural network according to the present invention. Hereinafter, the low data from the photon count detector 100 will be described with reference to FIGS. 7 to 9. The process of generating the correction data from which the virtual image is removed by acquiring the above will be described in detail.

First, referring to a method of removing an artifact of a CT image using a fuzzy neural network according to the present invention, in which the image compensator 230 is not included, the image acquisition unit 210 is a row obtained by tomography imaging an object from the photon count detector 100. Acquire data (S10). In the step (S10), the image acquisition unit 210 to detect the X-ray by a plurality of energy bands from the photon counting detector 100 to obtain the raw data for the object for each of the plurality of energy bands Can be.

The image correction unit 230 transmits the raw data including the virtual image to the image converting unit 220 (S20), and the image converting unit 220 converts the raw data to polar coordinates and detects the virtual image data having a line shape. (S30).

The image converting unit 220 transmits the detected virtual image data to the image processing unit 240 (S40), and the image processing unit 240 inputs the virtual image data to a fuzzy neural network and applies a fuzzy membership function. To form output data that converges to a predetermined value (S70).

More specifically, as shown in FIG. 8, the image converter 220 forms an input fuzzy set composed of a plurality of fuzzy data by applying a function of belonging to the virtual image data (S71), and by at least two virtual image data. After classifying fuzzy data by assigning degree of belonging to the formed fuzzy set according to distance from specific fuzzy data (S72), the classified fuzzy data selects inference data by fuzzy rule base and output fuzzy After the set is formed (S73), the output data may be formed through unpurging the output fuzzy set (S74).

The output data formed through this process is transmitted to the image conversion unit 220 (S80), and the image conversion unit 220 obtains correction data by converting the transmitted output data in Cartesian coordinates (S90). .

Meanwhile, as shown in FIG. 9, before the step S70, the image compensator 230 receives the virtual image data from the image converter 220 (S40), and corrects the virtual image data to correct the image processor 240. The corrected virtual data may be transmitted (S60).

At this time, the correction of the virtual image data may be performed in a manner of emphasizing the virtual image forming a line by gray level transformation of the virtual image data (S51), extracting a feature point of the virtual image data, and connecting the feature points. It can be made in such a way to grasp the shape by removing the pseudo-feature points (S52). In addition, a function for an arbitrary position is selected by using a high-order interpolation and a polynomial approximation function that selects a target pixel from the virtual data and takes an average value of other pixels surrounding the target pixel. The virtual data may be corrected through a hybrid interpolation method in which polynomial fitting-interpolation is obtained (S53).

All of the above-described correction method of the virtual image data may be performed, and at least one correction method may be performed.

Since the present invention described above can detect a more accurate virtual image from the image acquired from the photon counting detector 100 and can remove the detected virtual image accurately and conveniently, using this, a relatively wide chest, as well as a knee joint It is possible to make accurate diagnoses of various bodies such as bones of the hands, bones of the hands, sinuses and breasts, and the data accumulated through the fuzzy neural network can be utilized in various medical fields.

The invention can also be embodied as computer readable code on a computer readable recording medium. Computer-readable recording media include all kinds of recording devices that store data that can be read by a computer system.

Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disks, optical data storage devices, and the like, which are also implemented in the form of carrier waves (eg, transmission over the Internet). It also includes.

The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. And functional programs, codes and code segments for implementing the present invention can be easily inferred by programmers in the art to which the present invention belongs.

100: photon count detector
200: virtual image removal device
210: image acquisition unit
220: video conversion unit
230: Video Compensation
240: image processing unit

Claims (13)

Obtaining raw data including an artifact generated by tomography of an object from a photon counting detector;
Detecting the virtual data having a line shape by converting the raw data by polar coordination;
Emphasizing the virtual image forming a line by gray level transformation of the virtual image data;
Inputting grayscale-converted virtual image data into a fuzzy neural network and applying fuzzy membership function to form output data that converges to a predetermined value; And
And converting the output data into Cartesian coordinates to obtain correction data. The method of claim 1, wherein the forming of the output data comprises:
Applying an affiliation function to the virtual image data to form an input fuzzy set consisting of a plurality of fuzzy data;
Classifying the fuzzy data by assigning the input fuzzy set formed by the at least two virtual data to the degree of belonging according to the distance to the specific fuzzy data;
Selecting the inferred data by the classified fuzzy data by a fuzzy rule base to form an output fuzzy set; And
Forming output data through unpurging the output fuzzy set;
Virtual image removal method of the CT image using a fuzzy neural network, characterized in that it comprises a. delete The method of claim 1, wherein before forming the output data,
And extracting a feature point from the gray degree converted virtual image data, and identifying and removing a pseudo feature point by identifying a connection form of the extracted feature points. The method of claim 1, wherein before forming the output data,
A target pixel is selected from the grayscale transformed virtual image data, and a high-order interpolation that takes an average value of other pixels surrounding the target pixel at the target pixel and an approximation function composed of polynomials are used at arbitrary positions. The virtual image data removal method using the fuzzy neural network, characterized in that it further comprises the step of correcting the virtual image data through a hybrid Hid interpolation combined with polynomial fitting-interpolation to obtain a function value for. The method of claim 1, wherein the obtaining of the raw data comprises:
The method of removing a virtual image of a CT image using a fuzzy neural network, characterized by obtaining X-rays from the photon count detector by dividing the X-rays into a plurality of energy bands and obtaining raw data of the object for the plurality of energy bands. A computer-readable recording medium having recorded thereon a program for executing the method of any one of claims 1, 2, 4, 5, and 6 on a computer,
The recording medium is at least one of a ROM, RAM, CD-ROM, magnetic tape, floppy disk and optical data storage device. An image acquisition unit for acquiring raw data including an artifact generated by tomography imaging an object from a photon counting detector;
An image converter configured to detect the virtual data having a linear shape by converting the raw data into polar coordinates, and converting the output data formed using the virtual data into Cartesian coordinates to obtain correction data;
An image compensator for emphasizing the virtual image forming a line by gray level transformation of the virtual image data; And
A virtual image of a CT image using a fuzzy neural network, comprising: an image processor configured to input grayscale-converted virtual data to a fuzzy neural network and apply fuzzy membership function to form output data converged to a predetermined value Removal device. The method of claim 8,
The image processing unit forms an input fuzzy set composed of a plurality of fuzzy data by applying a belonging function to the virtual image data, and sets the degree of belonging to the input fuzzy set formed by at least two virtual image data according to a distance from a specific fuzzy data. Classify the fuzzy data, classify the fuzzy data, and classify the inferred data using a fuzzy rule base to form an output fuzzy set. Apparatus for removing virtual images of CT images using a fuzzy neural network, characterized in that to form a. delete The method of claim 8,
The video correction unit,
The virtual image removal apparatus of the CT image using a fuzzy neural network, characterized in that the feature points are extracted from the gray-scale converted virtual image data, and the pseudo feature points are selected and removed by grasping the connection form of the feature points. The method of claim 8,
The video correction unit,
A target pixel is selected from the grayscale transformed virtual image data, and a high-order interpolation that takes an average value of other pixels surrounding the target pixel at the target pixel and an approximation function composed of polynomials are used at arbitrary positions. A virtual image removal apparatus for CT images using a fuzzy neural network, characterized by correcting virtual data through a hybrid hierarchical interpolation method that combines polynomial fitting-interpolation. The method of claim 8,
The image acquisition unit detects an X-ray from a photon count detector for each of a plurality of energy bands, and acquires raw data of an object for each of the plurality of energy bands. .

Images