JP2678815B2 - Radiation image reader - Google Patents

Description

The present invention relates to a radiation image reading apparatus for reading a radiation image from a recording sheet such as a stimulable phosphor sheet on which radiation image information of a subject is recorded to obtain an image signal. It is about.

2. Description of the Related Art In various fields, reading a recorded radiation image to obtain an image signal, performing appropriate image processing on the image signal, and reproducing and recording the image have been performed in various fields. For example,
An X-ray image is recorded using an X-ray film having a low gamma value designed so as to be compatible with the subsequent image processing, and the X-ray image is read from the film on which the X-ray image is recorded and converted into an electric signal. By subjecting this electric signal (image signal) to image processing and then reproducing it as a visible image on a copy photograph or the like, a reproduced image with good image quality performance such as latitude, sharpness, and graininess can be obtained. There is
61-5193).

In addition, the applicant (X-ray, α-ray, β-ray,
A part of the radiation energy when irradiated with γ-rays, electron beams, ultraviolet rays, etc. is accumulated, and when it is irradiated with excitation light such as visible light after that, a stimulable phosphor that exhibits stimulated emission according to the accumulated energy ( The stimulable phosphor is used to temporarily record the radiation image information of a subject such as a human body into a sheet-shaped stimulable phosphor, and the stimulable phosphor sheet is scanned with excitation light such as a laser beam to emit a luminescent material. Generates exhausted light, photoelectrically reads the obtained stimulated emission light to obtain an image signal, and based on this image data the radiation image of the subject as a visible image on a recording material such as a photographic photosensitive material or a CRT. A radiation image recording / reproducing system for outputting has already been proposed (Japanese Patent Laid-Open No. 55-1242).
No. 9, 56-11395, 55-163472, 56-104645
No. 55-116340, etc.).

This system has the practical advantage of being able to record images over a very large radiation exposure area compared to conventional radiographic systems using silver halide photography. That is, in the case of the stimulable phosphor, it has been recognized that the amount of emitted light that is stimulated by excitation after accumulation is proportional to the radiation exposure amount over an extremely wide range. Even if fluctuates considerably, the amount of the stimulating light emitted from the stimulable phosphor sheet is read by the photoelectric conversion means with the reading gain set to an appropriate value and converted into an electric signal. By outputting a radiation image as a visible image on a recording material such as a photographic light-sensitive material or a display device such as a CRT using, a radiation image which is not affected by a change in radiation exposure can be obtained.

In the above system, the stimulable phosphor sheet is scanned with a low-level light beam in advance before the image signal is obtained by reading under the optimum reading condition according to the dose of the radiation applied to the stimulable phosphor sheet. Pre-reading is performed to read the outline of the radiation image recorded on this sheet, the pre-reading image signal obtained by this pre-reading is analyzed, and then the sheet is irradiated with a high-level light beam and scanned, and this radiation image is read. There is a system configured to perform a main reading in which an image signal is obtained by reading under optimum reading conditions.

Here, the reading condition is a general term for various conditions that affect the relationship between the amount of stimulated emission light in reading and the output of the reading device, and can be divided into latitude Gp and sensitivity Sk. The latitude Gp corresponds to the light intensity ratio of the strongest stimulated emission light that is correctly converted to an image signal to the weakest stimulated emission light that is correctly converted to an image signal, and the sensitivity Sk is the sensitivity of a predetermined light amount. This corresponds to the photoelectric conversion rate at which level the stimulated emission is used as the image signal.

Further, the high level / low level of the light beam means that the energy of the light beam applied per unit area of the sheet is large / small or the energy of the stimulated emission light emitted from the sheet is the light beam. (Having a wavelength sensitivity distribution) means the magnitude of the weighted energy after weighting the energy of the light beam irradiated per unit area of the sheet by the wavelength sensitivity. Methods for changing the beam level include using a light beam of a different wavelength, changing the intensity of the light beam emitted from a laser light source, etc., and inserting or removing an ND filter or the like in the optical path of the light beam. How to change the beam intensity, how to change the scanning density by changing the beam diameter of the light beam, how to change the scanning speed, etc.
Various known methods can be used.

In addition, regardless of whether the system performs the pre-reading or the system which does not perform the pre-reading, the obtained image signal (including the pre-reading image signal) is analyzed, and an optimum image processing condition when performing image processing on the image signal is determined. Some systems let you decide. Here, the image processing conditions can be divided into latitude Gp and sensitivity Sk.
Gp corresponds to the conversion ratio of the span of the image signal, which is how many digits of the image signal before the image processing is to be all the digits (for example, 4 digits) of the image signal after the image processing, and the sensitivity Sk is the sensitivity of the image processing. It corresponds to the shift amount of the value of the signal indicating which value the image signal having the predetermined value before applying is converted into. The method for determining the optimum image processing conditions based on this image signal is not limited to a system using a stimulable phosphor sheet, and obtains an image signal from a radiation image recorded on a recording sheet such as a conventional X-ray film. It is also applied to the system.

A calculation for obtaining the reading condition and / or the image processing condition based on the image signal (including the preread image signal) (hereinafter, EDR
Call. ) Indicates the number of radiographic images classified according to imaging conditions such as the imaging region of the subject (head, chest, abdomen, etc. when the subject is a human body) and imaging method (normal imaging, contrast imaging, enlarged imaging, etc.). From the result of statistically processing a large number of image signals obtained from each of them, an algorithm is determined for each photographing condition, and EDR is automatically performed widely. This algorithm is usually defined in terms of how to obtain a histogram of the image signal and analyze this histogram.

(Problems to be Solved by the Invention) However, the imaging region of the subject and the imaging method are the same, the shape of the histogram is also very similar, and the region of interest to be observed in the radiographic image is also the same. However, there are cases in which it is necessary to obtain different reading conditions, and in this case, there is a problem that it is not possible to deal with the algorithm that has been conventionally determined. For example, in an X-ray image obtained by normal X-ray imaging centering on the shoulder of the human body, a vertebral body may or may not be included in the X-ray image due to a slight shift of the subject at the time of imaging.
Alternatively, the area of the lung field may be slightly larger or smaller. In this case, the presence or absence of the thruster and the area of the lung field may be slightly wide or narrow. In this case, the histogram of the image signal obtained by reading the X-ray image has substantially the same shape regardless of the presence or absence of the vertebral body and the size of the lung field. However, when the vertebral body is not desired to be imaged in the X-ray image or the lung field is narrowly imaged, it is necessary to obtain the reading condition or the like corresponding to the image signal having a smaller value on the histogram. When the vertebral body is imaged in the line image or the lung field is widely imaged, it is necessary to obtain the reading condition or the like corresponding to the image signal having a larger value on the histogram.

Since such a distinction cannot be made by the conventional algorithm, for example, a wide reading condition of latitude Gp is set so as to have a proper reading condition to some extent regardless of the presence or absence of the vertebral body or the lung field and whether it is wide or narrow. . In this case, it becomes necessary to represent an image having a wide signal latitude Gp composed of a predetermined number of bits per pixel, so the density resolution of the image decreases, and a visible image is reproduced and output based on this image signal. When this is done, the area required for observation does not have an appropriate density (including the brightness when a visible image is reproduced and displayed on a CRT display or the like) on the visible image, and it may become whitish or blackish. There was a problem.

In view of the above circumstances, it is an object of the present invention to provide a radiographic image reading apparatus that can obtain appropriate reading conditions and the like even if the imaging region and the imaging method are the same and the histograms are similar. Is.

(Means for Solving the Problems) A first radiographic image reading apparatus of the present invention for solving the above-mentioned object relates to a radiographic image reading apparatus that performs prefetching using a stimulable phosphor sheet. That is, the first radiographic image reading device of the present invention, the stimulable phosphor sheet on which the radiographic image is recorded is irradiated with excitation light, and the stimulated emission light emitted from the stimulable phosphor sheet by the irradiation of the excitation light. Pre-reading means for obtaining a pre-reading image signal representing the radiation image by irradiating the stimulable phosphor sheet again with excitation light, and the stimulated emission emitted from the stimulable phosphor sheet by the irradiation of the excitation light. Main reading means for reading light to obtain a main reading image signal representing the radiation image, reading conditions for obtaining the main reading image signal based on the histogram obtained by obtaining a histogram of the preread image signal, and / or obtained Latitude calculation means for obtaining a latitude among image processing conditions when performing image processing on the main reading image signal, and the preread image signal as input, and the reading condition and / or the previous It is characterized in that it is provided with a sensitivity calculation means composed of a neural network which outputs the sensitivity among the image processing conditions.

The second radiation image reading device of the present invention is not limited to the stimulable phosphor sheet, and also relates to a radiation image reading device that does not perform pre-reading. That is, the second radiographic image reading device of the present invention is a reading unit that reads the radiographic image from a recording sheet on which a radiographic image is recorded to obtain an image signal representing the radiographic image, obtains a histogram of the image signal, and displays the histogram in the histogram. A latitude calculation means for obtaining a latitude among image processing conditions when performing image processing on the image signal based on the
And a sensitivity calculation unit including a neural network that receives the image signal as an input and outputs the sensitivity among the image processing conditions.

(Operation) When the imaged region and the imaged method are the same and the region of interest to be observed is the same, the latitudes Gp are the same and only the sensitivities Sk are different even if the position of the subject shifts during imaging.

Therefore, in the present invention, the latitude Gp is obtained by obtaining the histogram of the image signal (including the preread image signal), and the sensitivity Sk is obtained directly from the radiation image using the neural network. It is a thing.

Here, the neural network is a method having a completely different concept from the conventional methods, which can be used as a recognition method for inputting predetermined information, performing some recognition, and outputting the recognition result. By inputting information (teacher signal) indicating whether the output signal output when the input signal is given is a correct signal or an incorrect signal, the weight of the connection between the units inside the neural network (synapse connection) The error backpropagation learning (backpropagation) function of correcting (weight of) is provided. By repeatedly'learning 'this neural network, the probability of outputting a correct answer when a new signal is input can be increased (for example, “DERumelhart, GE Hinton and RJWilli
ams: Learning representations by back-propagating
errors, Nature, 323-9, 533-356, 1986a "," Hideki Aso: Back Propagation Computrol No.24 53-6
0 "," Kazuyuki Aihara Neural Computer Tokyo Denki University Press ").

Therefore, by letting the neural network perform'learning 'so that the correct sensitivity Sk is output when the above image signals (including the pre-read image signal) are input, if there is a position shift of the subject during shooting, this position shift will occur. The correct sensitivity Sk is obtained by recognizing the above.

In this way, the latitude Gp is obtained by the conventional histogram analysis, and the sensitivity Sk is obtained by using the neural network.Therefore, there is a position shift of the subject at the time of shooting, and it is not possible to obtain it by the conventional histogram analysis alone. Reading conditions are required.

Here, the neural network can be configured so that not only the sensitivity Sk but also the latitude Gp is obtained by using the neural network.

However, if a neural network is constructed so as to obtain both the sensitivity Sk and the latitude Gp, the neural network becomes very complicated, and it is a huge storage device for storing the weight of the coupling between the units constituting this neural network. It is necessary to prepare one with a sufficient memory capacity, and a very enormous number of times of'learning '.
Therefore, it takes a long time to obtain the sensitivity Sk and the latitude Gp after the image signal (including the preread image signal) is input. Compared with the configuration of obtaining both the sensitivity Sk and the latitude Gp using the neural network of the present invention, the configuration is simplified as a whole by using both the histogram analysis and the neural network, and the memory having a smaller storage capacity is obtained. It has an advantage that a medium can be prepared, the sensitivity Sk and the latitude Gp can be obtained in a short time, and the number of learnings of the neural network can be small.

(Examples) Hereinafter, examples of the present invention will be described. Here, an example in which the above-described stimulable phosphor sheet is used will be described.

 FIG. 1 is a schematic diagram of an example of an X-ray imaging apparatus.

The X-ray source 2 of the X-ray imaging apparatus 1 irradiates the human body 4 with X-rays 3, and the X-rays 3a that have passed through the human body 4 are radiated to the stimulable phosphor sheet 11 to transmit X-rays of the human body. A line image is accumulated and recorded on the stimulable phosphor sheet 11.

2A and 2B are diagrams schematically showing an example of an X-ray image accumulated and recorded on the stimulable phosphor sheet.

2A and 2B both show X-ray images of the right shoulder, the vertebral body 5 is photographed in FIG. 2A, and the lung field 6 is also widely photographed. On the other hand, in FIG. 2B, the vertebral body is not imaged, and the lung field 6 is also imaged only in a small area. Although only the X-ray image of the right shoulder is shown here, this is only an example, and the head,
Images of various parts such as the chest and abdomen are taken.

FIG. 3 is a perspective view showing an example of an X-ray image reading apparatus which is an embodiment of the radiation image reading apparatus of the present invention. This device is a device for performing pre-reading using the above-mentioned stimulable phosphor sheet.

The stimulable phosphor sheet 11 on which an X-ray image is recorded in the X-ray imaging apparatus 1 shown in FIG. 1 is first scanned with a weak light beam and a part of the radiation energy accumulated in the stimulable phosphor sheet 11 is scanned. It is set at a predetermined position of the pre-reading means 100 for emitting only the pre-reading. The stimulable phosphor sheet 11 set at this predetermined position is transferred by a sheet conveying means 13 such as an endless belt driven by a motor 12,
It is transported (sub-scanning) in the direction of arrow Y. On the other hand, a weak light beam 15 emitted from the laser light source 14 is reflected and deflected by a rotating polygon mirror 16 driven by a motor 23 and rotating at a high speed in the direction of an arrow, passes through a focusing lens 17 such as an fθ lens, and then an optical path by a mirror 18. Changeable storage phosphor sheet 11
And an arrow X substantially perpendicular to the sub-scanning direction (arrow Y direction)
Main scan in the direction. From the portion of the stimulable phosphor sheet 11 where the light beam 15 is irradiated, accumulated and recorded X is recorded.
A quantity of stimulated emission light 19 corresponding to the line image information is diverged, and this stimulated emission light 19 is guided by a light guide 20 and photoelectrically detected by a photomultiplier (photoelectron image multiplier) 21. The light guide 20 is formed by molding a conductive material such as an acrylic plate, and has a linear incident end face 20.
a is arranged so as to extend along the main scanning line on the stimulable phosphor sheet 11, and the light receiving surface of the photomultiplier 21 is coupled to the emission end surface 20b formed in an annular shape. Photostimulated luminescent light 1 that has entered the light guide 20 from the incident end face 20a.
9 proceeds by repeating total reflection inside the light guide 20,
The photomultiplier 21 emits the light from the emission end face 20b and receives the photostimulated luminescence light 19 representing the radiation image, and the photomultiplier 21 converts the amount of the stimulated emission light 19 into an electric signal.

The analog output signal S output from the photomultiplier 21 is logarithmically amplified by a logarithmic amplifier 26, and the A / D converter 2
Digitization is performed in step 7 to obtain a pre-read image signal Sp. The signal level of this pre-read image signal Sp is a stimulable phosphor sheet.
It is compared with the logarithm of the amount of stimulated emission light emitted from each of the 11 pixels.

In the pre-reading, the reading conditions, that is, the voltage value applied to the photomultiplier 21 and the amplification factor of the logarithmic amplifier 26 are determined so that the radiation energy accumulated in the stimulable phosphor sheet 11 can be read over a wide region. Has been.

The read-ahead image signal Sp obtained is stored in the computer system 40.
Is input to This computer system 40 includes an example of each of the latitude calculating means and the sensitivity calculating means according to the present invention, and is a main body portion 41 in which a CPU and an internal memory are incorporated, and a floppy disk is inserted and driven as an auxiliary memory. A drive unit 42, a keyboard 43 for an operator to input necessary instructions and the like to the computer system 40, and a CRT display 44 for displaying necessary information.

FIG. 4 is a block diagram illustrating how to determine the reading conditions performed in the computer system 40.

When the pre-reading image signal Sp is input into the computer system 40, the pre-reading image signal Sp is subjected to histogram analysis and also input to the neural network, and the latitude Gp of the reading conditions is obtained by the histogram analysis, and The sensitivity Sk is output from the neural network.

The computer system 40 includes subject parts (head, neck, chest, abdomen, etc.) and their orientations (front, front,
Histogram for obtaining the latitude Gp suitable for each of these conditions, which differs depending on the average value of the image signal (average value of the radiation energy accumulated in the stimulable phosphor sheet), etc.). An analysis algorithm and a neural network for obtaining the sensitivity Sk suitable for each of these conditions (referred to as a coefficient representing the connection weight of each neuron that constitutes the neural network, which may be simply referred to as a neural network hereinafter) are stored. When the pre-read image signal Sp is input to this computer system, the corresponding histogram analysis algorithm and the corresponding neural network are read from the memory in this computer system, and based on these histogram analysis algorithm and neural network. Each latitude Gp and sensitivity Sk during real reading is obtained, this sensitivity Sk, according latitude Gp e.g. photomultiplier 21 '
The voltage value to be applied to the amplifier and the amplification factor of the logarithmic amplifier 26 'are controlled.

The pre-reading stimulable phosphor sheet 11 'is set at a predetermined position of the main reading means 100', and the sheet 11 'is scanned by the light beam 15' which is stronger than the light beam used for the pre-reading. An image signal can be obtained under the reading conditions (sensitivity Sk, latitude Gp) obtained based on Sp, but since the configuration of the main reading means 100 'is substantially the same as the configuration of the above-mentioned pre-reading means 100, the pre-reading means 100 The numbers used in the pre-reading means 100 are shown with a dash added to the components corresponding to each of the components, and the description thereof will be omitted.

The image signal S _Q obtained by being digitized by the A / D converter 27 ′ is input to the computer system 40 again. In the computer system 40, the image signal S _Q is subjected to appropriate image processing, the image signal subjected to this image processing is sent to a reproducing device (not shown), and the reproducing device reproduces and displays an X-ray image based on this image signal. To be done.

Here, an example of an algorithm for obtaining the latitude Gp by histogram analysis will be described.

FIG. 5 is a diagram showing an outline of a histogram of the pre-read image signal Sp obtained by reading the X-ray images shown in FIGS. 2A and 2B. Here, even when the vertebral body is photographed and the lung field of a large area is photographed as shown in FIG. 2A,
Even if the vertebral body is not imaged and the lung field of a small area is imaged as shown in FIG. 2B, the histogram is almost similar, and therefore 1 here.
It is represented by two histograms.

The pre-read image signal Sp, which is obtained by reading the stimulated emission light emitted from the stimulable phosphor sheet in the pre-reading, is taken on the horizontal axis (logarithmic axis) and is proportional to the amount of the stimulated emission light. The frequency of appearance of each value of Sp is plotted on the vertical axis (upper), and the image signal obtained by the main reading is plotted on the lower side of the vertical axis (logarithmic axis). At this time, the histogram obtained from the pre-read image signal Sp is divided into two peaks A and B as shown in the figure. Here, in the mountain B, the direct X-ray portion 7 in which the X-ray 3 does not pass through the subject 4 and is directly applied to the stimulable phosphor sheet 11
It corresponds to (see FIGS. 2A and 2B) and is an unnecessary part in reproducing an image. The mountain A is a mountain corresponding to the subject part which is irradiated on the stimulable phosphor sheet 11 after the X-ray 3 has passed through the subject 4, but the area required for image reproduction is taken as shown in FIG. 2A. Unlike the case where the image is taken as shown in FIG. 2B, the areas 8 and 9 correspond to each other. Here, in order to correctly obtain the reading conditions (both the sensitivity Sk and the latitude Gp) at the time of the main reading for the X-ray image shown in FIG. 2A, the minimum pre-read image signal Sp1 in the area 8 is the main reading. Corresponding to the minimum image signal S _Qmin obtained, the maximum pre-read image signal Sp2 of the area 8 corresponds to the maximum image signal S _Qmax obtained in the main reading, that is, to the straight line G1 shown in FIG. Read conditions (sensitivity Sk and latitude
Gp) must be calculated, and similarly, Xp shown in FIG.
In order to correctly obtain the reading condition for the main reading for the line image, it is necessary to obtain the reading condition so as to correspond to the straight line G2. However, since it is impossible to distinguish from the histogram whether the image was taken as shown in FIG. 2A or the image shown in FIG. 2B, the image of FIG. 2A and the image of FIG. It is impossible to obtain the reading condition along the line and the reading condition along the straight line G2.

Therefore, conventionally, it is impossible to obtain the reading condition along the straight lines G1 and G2, so that the value of the preread image signal Sp increases from the position of the minimum value Sp1 of the preread image signal Sp at the threshold value T, for example. Search in the direction (corresponding to the chain line in the figure),
The first rising point a and the next falling point b are found, and the minimum value and the maximum value of the range D sandwiched between these two points a and b are the minimum values S of the image signal S _Q obtained during the actual reading. _The reading conditions were required to be _Qmin and the maximum value S _Qmax , that is, along the straight line G3. In this case, as described above, since the latitude is wide, there arises a problem that an image with low density resolution is obtained. Here, the sensitivity Sk corresponds to the positions of the straight lines G1, G2, and G3 in the horizontal direction of the figure (meaning that the sensitivity Sk is higher toward the left), and the latitude Gp corresponds to the inclination of these straight lines (standing It means that the smaller the latitude Gp, the smaller it is). Comparing the straight line G1 with the straight line G2, the sensitivities Sk (positions in the left-right direction) are significantly different from each other, but the latitudes (tilts) are almost the same. It is empirically known that when the regions of interest are the same, the latitude Gp has almost the same value even if the position of the subject shifts during shooting. Therefore, in this embodiment, a predetermined ratio (for example, 1/3) of the range D sandwiched between the two points a and b obtained as described above.
Is calculated as the latitude Gp, and the sensitivity Sk is left to the neural network described below.

Figure 6 shows error backpropagation learning (backpropagation).
It is a figure showing an example of a neural network provided with a function. Backpropagation learning (backpropagation)
As mentioned above, is a "learning" algorithm that corrects the weight of the sequential connection (the weight of the synaptic connection) from the output side to the input side by comparing the output of the neural network with the correct answer (teacher signal). Say.

As shown in the figure, the first part of this neural network is
The layer (input layer), the second layer (intermediate layer), and the third layer (output layer) are composed of n ₁ , n ₂ , and 1 units (neurons), respectively. Signals F ₁ and F input to the first layer (input layer)
₂ , ..., F _n1 is the pre-read image signal Sp corresponding to each pixel after thinning out the X-ray image, and the output ▲ y ³ ₁ ▼ from the third layer (output layer) is for the main reading. This signal corresponds to the sensitivity Sk of.
The i-th unit of the k-th layer is represented by ▲ u ^k _i ▼, and the unit ▲ u
^k _i ▲ each input to the ▼ x ^k _i ▼, each output _{^{▲ y k i ▼, ▲ u}} k i
The weight of the connection from ▼ to ▲ u ^{k + 1} _j ▼ is ▲ W ^{k k + 1} _ij ▼, and each unit ▲ u ^k _j ▼ has the same characteristic function. Shall be provided. In this case, each unit ▲ u ^k _j ▼ input ▲ x ^k _j ▼, output ▲ y ^k _j ▼ is Becomes However, each unit constituting the input layer ▲ u ¹ _i ▼
Each input F ₁ , F ₂ , ..., F _n1 to (i = 1,2, ..., n ₁ ) is not weighted and each unit ▲ u ¹ _i ▼ (i = 1,2, ...,
n ₁ ). Input n ₁ signals F ₁ , F ₂ , ..., F _n1
Is transmitted to the final output ▲ y ¹ ₃ ▼ while being weighted by the weight of each coupling ▲ W ^{kk + 1} _ij ▼, whereby the sensitivity Sk among the reading conditions at the time of actual reading is obtained.

Here, a method of determining the weight ▲ W ^{kk + 1} _ij ▼ of each connection will be described. First, the weight of each connection by random numbers ▲ W ^{kk + 1}
The initial value of _ij ▼ is given. At this time, even if the inputs F _{1 to} F _n1 fluctuate to the maximum, it is possible to limit the range of the random numbers so that the output ▲ y ³ ₁ ▼ will be a value within a predetermined range or a value close thereto. preferable.

A large number of stimulable phosphor sheets on which X-ray images with known optimum reading conditions (sensitivity Sk) are recorded are read as described above, and the pre-read image signals Sp thus obtained are thinned out to obtain the above n ₁ The inputs F ₁ , F ₂ , ..., F _n1 are obtained. This n ₁
Inputs _{F 1, F 2, ...,} F n1 is input to the neural network shown in FIG. 6, each unit ▲ u ^k _i ▼ output ▲ y ^k _i ▼
Is monitored.

When each output ▲ y ^k _i ▼ is obtained, the final output ▲ y ³ ₁ ▼ and the teacher signal as a correct reading condition for this image Square error with Is required. In order to minimize the double error E _1, the weight ▲ W ^{kk + 1} _ij ▼ of each connection is corrected as follows.

To minimize the double error E ₁ , this E ₁ is ▲ W ^{kk + 1} _ij ▼
Because it is a function of As described above, the weight ▲ W ^{kk + 1} _ij ▼ of each connection is corrected. Here, η is a coefficient called a learning coefficient.

here, And from equation (2) Therefore, equation (6) is Becomes

Here, from equation (4), When this equation (8) is transformed using equation (3), Here, from the equation (1), f ′ (x) = f (x) (1-f (x)) (10) Becomes

Putting k = 2 in the equation (7) and substituting the equations (9) and (11) into the equation (7), Substituting equation (12) into equation (5), Becomes According to this equation (13), ▲ W ^{2 3} _{i 1} ▼ (i =
The weight of each connection of 1,2, ..., n ₁ ) is modified.

next, Therefore, substituting equations (2) and (3) into equation (14), From equation (10), Therefore, the equation (16) and the equations (9) and (11) are (1
Substituting in 5) formula, Putting k = 1 in the equation (7) and substituting the equation (17) into the equation (7), Substituting equation (18) into equation (5), we put k = 1, And ▲ W ^{2 3} _{i 1} ▼ (i = 1,
2, ..., n ₁ ) is substituted into this equation (19), and ▲ W ^{1 2} _ij ▼
(I = 1,2, ..., n ₁ ; j = 1,2, ..., n ₂ ) are modified.

Theoretically, using the equations (13) and (19), the learning coefficient η is set to be sufficiently small and the number of times of learning is set to be sufficiently large so that the weight ▲ W ^{k k + 1} _ij ▼ of each connection is predetermined. However, it is not realistic to make the learning coefficient η too small because the learning progress is delayed. On the other hand, the learning coefficient η
When is large, learning may oscillate (the weight of the above coupling does not converge to a predetermined value). Therefore, in practice, an inertial term such as a time equation is added to the correction amount of the coupling weight to suppress vibration, and the learning coefficient η is set to a relatively large value. (E.g., DERumelhart, GEHinton and RJWil
liams: Learning internal representations by error p
ropagation In Parallel Distributed Processing, Volu
me 1, JLMcClelland, DERumelhart and The PDP Rese
arch Group, MIT Press, 1986b '') However, ▲ ΔW ^{k k + 1} _ij ▼ (t) is ^obtained by changing the weight ▲ W ^{k k + 1} _ij ▼ of the connection before correction from the weight ▲ W ^{k k + 1} _ij ▼ of the correction in the t-th learning. Indicates the amount of correction subtracted. Α is a coefficient called an inertia term.

As the inertia term α and the learning coefficient η, for example, α = 0.9 η = 0.
Perform each coupling weight ▲ W ^{k k + 1} _ij ▼ modification using 25 (learning) e.g. 200,000 times, then, the coupling of the weight ▲ W ^{k k + 1} _ij ▼ is fixed to the final value It Here, in order to continue the learning as described below, after the system including the recognition device for obtaining the reading condition using this neural network is installed in the user,
The final value here means, for example, the final value at the initial startup stage of the user. At the end of this learning, the output {circle around (y) ^} ₁ becomes a signal that almost correctly represents the sensitivity Sk at the time of actual reading.

A neural network for which learning has been completed as described above is prepared, and different sensitivities Sk are obtained for the X-ray image of FIG. 2A and the X-ray image of FIG. 2B, for example, using this neural network. The above-mentioned actual reading is performed using the sensitivity Sk obtained by this neural network and the latitude Gp obtained by the above-mentioned histogram analysis.

In the above embodiment, the apparatus for determining the reading condition in the main reading by using the histogram analysis and the neural network has been described, but in the main reading, the reading should be performed under the predetermined reading condition regardless of the pre-read image signal Sp. age,
With histogram analysis and neural networks,
Determining the image processing conditions for performing image processing on the image signal S _Q, the image processing conditions may be subjected to image processing to the read image signal S _Q using, also, the read condition and the image processing Both conditions may be determined.

In the embodiment described with reference to FIG. 3, the pre-reading means 100 and the main reading means 100 'are separately configured, but as described above, the configurations of the pre-reading means 100 and the main reading means 100' are omitted. Since they are the same, the pre-reading means 100 and the real-reading means 100 'may be combined. In this case, after pre-reading by scanning with a weak light beam, the storage phosphor sheet 11 is backed once,
This time, the main reading may be performed again by scanning with a strong light beam.

When the pre-reading means and the main reading means are used together, it is necessary to switch the light beam intensity between the case of pre-reading and the case of main reading. ND on the optical path of the light beam
It is possible to use various known methods such as a method of switching the light intensity by inserting and removing a filter, a method of changing the beam diameter of the light beam, a method of switching the speed of the main scanning and the speed of the sub-scanning. it can.

FIG. 7 is a perspective view showing another embodiment of the radiation image reading apparatus of the present invention. This embodiment is similar to the above-mentioned embodiment in that the stimulable phosphor sheet is used, but it is a system in which pre-reading is not performed.

In this embodiment, the reading means 100 'is substantially the same as the main reading means 100' in the embodiment shown in FIG.
Corresponding components are assigned the same reference numerals and explanations thereof are omitted.

The image signal S _Q output from the A / D converter 27 'is input to the computer system 40'. In the computer system 40 ′, in the same manner as in the above-described embodiment, the latitude among the image processing conditions for performing image processing on the image signal S _Q so that a visible image having appropriate density and contrast is output. Is obtained by histogram analysis, and the sensitivity is obtained by a neural network. The image signal subjected to such image processing is sent to an image output device (not shown), and the image output device outputs a hard copy of the radiation image based on the sent image signal.

Although the three-layer structure neural network has been described in the above embodiment (see FIG. 6), it goes without saying that the present invention is not limited to the use of the three-layer structure neural network. Moreover, it goes without saying that the number of units u ^k _i constituting each layer, the number of input points, and the like can be arbitrarily configured according to the purpose.

Further, although each of the above embodiments is an example of an apparatus using a stimulable phosphor sheet, it goes without saying that the present invention can also be applied to an apparatus using a conventional X-ray film or the like.

(Effects of the Invention) As described in detail above, the radiation image reading apparatus of the present invention includes the latitude calculation means for obtaining the latitude by the histogram analysis and the sensitivity calculation means for obtaining the sensitivity using the neural network. It is possible to obtain correct reading conditions and the like which were impossible only by the conventional histogram analysis. Further, the configuration can be simplified as a whole as compared with the case where both the sensitivity and the latitude are obtained by a neural network.

[Brief description of the drawings]

FIG. 1 is a schematic diagram of an example of an X-ray imaging apparatus, and FIGS. 2A and 2B are diagrams schematically showing an example of an X-ray image accumulated and recorded on a stimulable phosphor sheet. FIG. 4 is a perspective view showing an example of an X-ray image reading apparatus which is an embodiment of the radiation image reading apparatus of the present invention, FIG. 4 is a block diagram showing how to obtain reading conditions, and FIG. FIG. 6 is a diagram showing an example of a histogram of a preread image signal, FIG. 6 is a diagram showing an example of a neural network having an error backpropagation learning function, and FIG. 7 is another example of the radiation image reading apparatus of the present invention. Perspective view showing an embodiment, 1 ... X-ray imaging device, 2 ... X-ray source 11,11 '... Accumulable phosphor sheet 19,19' ... Stimulated emission light 21,21 '... Photo Multiplier 26,26 ′ …… Logarithmic amplifier 27,27 ′ …… A / D converter 40,40 ′ …… Computer system 100 …… Pre-reading means, 100 ′ …… Reader in stage

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display location H04N 7/18 A61B 6/00 303J

Claims (2)

(57) [Claims] 1. A radiation image is displayed by irradiating a stimulable phosphor sheet on which a radiation image is recorded with excitation light and reading the stimulated emission light emitted from the stimulable phosphor sheet by the irradiation of this excitation light. Pre-reading means for obtaining a pre-reading image signal, irradiating the stimulable phosphor sheet again with excitation light and reading the stimulated emission light emitted from the stimulable phosphor sheet by irradiation of this excitation light to obtain the radiation image Main reading means for obtaining a real reading image signal, a reading condition for obtaining the main reading image signal based on the histogram obtained by obtaining a histogram of the preread image signal, and / or
Alternatively, a latitude calculation unit that obtains a latitude among the image processing conditions when performing image processing on the obtained main reading image signal, and the preread image signal as an input, and the reading condition and / or the image processing condition A radiation image reading apparatus, comprising a sensitivity calculation means composed of a neural network that outputs sensitivity. 2. A reading means for reading the radiation image from a recording sheet on which the radiation image is recorded to obtain an image signal representing the radiation image, obtaining a histogram of the image signal, and performing image processing on the image signal based on the histogram. Latitude calculation means for obtaining a latitude among the image processing conditions when performing the image processing, and sensitivity calculation means composed of a neural network that receives the image signal and outputs the sensitivity among the image processing conditions. Radiation image reading device.