JP3005540B1 - Calculation image creation method - Google Patents

Abstract

Kind Code: A1 A thermal image in medical thermography is affected by many physical and physiological factors. Therefore, in diagnosing the thermal image, factors unrelated to a disease state need to be removed. In the prior art, this operation was not sufficient. Therefore, an object of the present invention is to improve a method of extracting pathophysiological information from a thermal image and improve a diagnostic result of thermography. SOLUTION: In the medical thermography, the skin temperature recovery characteristic after cooling load is approximated by a theoretical expression so as to faithfully reflect the pathophysiological function, and a plurality of parameters of this function are obtained from a time-series thermal image. The accuracy of the thermographic diagnosis is improved by obtaining the values for each of the pixels constituting these and further imaging each parameter value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to calculation in load thermography which can improve a diagnostic system in medical thermography diagnosis using load thermography measured by applying a thermal load such as cooling or heating to a subject. Related to image creation method.

[0002]

2. Description of the Related Art In general, thermography is known as a test method in which the temperature of the skin is measured by detecting infrared rays emitted from the body surface, the distribution of the skin is displayed on an image, and the image is used for diagnosis. The medical thermal image (temperature distribution image of the skin surface measured by sensing infrared rays) used in the thermography is affected by many physical and physiological factors. Therefore, in order to increase the accuracy rate in thermography, it is necessary to remove factors unrelated to the pathological condition. On the other hand, it has been confirmed that the skin temperature recovery characteristic after cooling load can be approximated by a single exponential function, and that the curve reflects a subcutaneous disease state. Then, the skin temperature recovery characteristic of the pixel (i, j) with respect to the time t after the cooling load of the thermal image data composed of n rows and m columns after the cooling load is:
It is shown by the following equation.

[0003]

(Equation 1)

Therefore, as one of the operations for removing factors unrelated to the pathological condition from the medical thermal image, time-series thermal image data obtained by measuring the temperature distribution on the skin surface of the subject over time, From (1), the parameter μ (i, j) of the above formula is obtained for each pixel, and the obtained value is imaged (by applying a different color to the pixel for each parameter value), thereby reflecting the pathophysiological function. (Calculated image) can be considered. The calculated image obtained in this manner is named a μ-value image, and has been applied to breast thermography diagnosis to obtain useful results.

[0005]

By the way, in the above-mentioned conventional method, it is often recognized that the temperature data immediately after the end of the cooling load does not follow the above-mentioned approximate expression. Therefore,
In order to more accurately reflect the pathophysiology in the μ-value image, it is necessary to approximate the skin temperature recovery curve with a more detailed theoretical formula. That is, the conventional method is not always sufficient for removing factors unrelated to a disease state included in a thermal image in medical thermography.

An object of the present invention is to provide a calculation image creation method in load thermography that can improve the method of extracting pathophysiological information from the above-mentioned time-series thermal images.

[0007]

According to a first aspect of the present invention, there is provided:
The calculation image creation method is to apply a load such as cooling, heating, etc.
Shows the temperature distribution on the subject's skin surface and its temporal change
A time-series temperature distribution image composed of n rows and m columns of pixels and a negative
Create a calculated image from the formula that approximates the skin temperature recovery characteristics after loading.
In the calculation image creation method to be performed, a temperature that changes with the passage of time for each pixel obtained from the time-series temperature distribution image, and a formula that approximates the skin temperature recovery characteristic after loading with the sum of a plurality of exponential functions, It is an object of the present invention to calculate the values of a plurality of parameters in the above formula for each pixel, and then create a calculation image indicating the values of the parameters for each pixel in different colors for each of the plurality of parameters. And

According to the above configuration, the approximate expression of the skin temperature recovery characteristic (skin temperature recovery curve) after a load such as cooling or heating is:
Conventionally, instead of using an expression having a single exponential function, an expression represented by the sum of a plurality of exponential functions that can be approximated by a skin temperature recovery curve is used. Factors unrelated to the disease state can be removed more reliably. Note that the approximate expression represented by the sum of a plurality of exponential functions basically includes a plurality of parameters.
By creating a calculation image for each parameter, it is possible to determine a morbid state or the like for each calculation image, and
Finally, it is possible to make a comprehensive judgment using a plurality of calculation images, and it is possible to improve the accuracy of medical thermography diagnosis.

[0009]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a configuration of a load thermography according to an embodiment of the present invention; In the calculation image creation method in the load thermography of this example, first, an infrared video camera (medical infrared imaging device) and a thermal image (an image showing the temperature distribution on the skin surface) used for medical thermography diagnosis by an image processing device or the like are used. The subject is photographed using a known device that can be obtained.

At this time, a temperature load such as cooling or heating is applied to the test subject, and the test is performed until the changed test subject's skin temperature returns to almost the original steady state, that is, until the skin temperature recovers. Shooting almost continuously, as shown in FIG.
Time-series thermal image data (t0 to
tn). In FIG. 1, the subjects are n rows,
Although shown as an image composed of m columns of pixels, in reality, each pixel is given a color corresponding to the skin temperature value (a color that differs depending on the temperature value (including differences in shading)). It has become. The load is cooling in this example, but may be heating. Further, n and m are integer values, and their values are not particularly limited, and are generally determined by the above-described apparatus for obtaining a thermal image. In addition, an arbitrary pixel among the pixels in the n-th row and the m-th column is defined as a pixel in the i-th row and the j-th column.

Next, using the time-series image data and the approximate expression of the skin temperature recovery curve when a subject is subjected to a thermal load, the values of the parameters of the approximate expression are determined. Here, the skin temperature recovery characteristic after the cooling load is approximated by the following equation obtained by introducing a theoretical analysis result. That is, the time immediately after the end of the cooling load is t, and the pixel (i, j)
Assuming that the temperature at time t is T (i, j, t), the following expression is used as the above approximate expression that approximates the skin temperature recovery characteristic of the pixel (i, j).

[0012]

(Equation 2)

Each parameter in the above approximation formulas (formulas (2) and (3)) includes information such as blood flow volume and thermal conductivity of skin and other tissues. More specifically, the parameters μ1 and μ2 are related factors such as blood flow, blood specific heat, tissue density, tissue specific heat, tissue thermal conductivity, and heat transfer coefficient. The parameters α and β are related factors such as tissue metabolic heat, nuclear temperature, blood temperature, environmental temperature, blood flow, blood specific heat, tissue density, tissue specific heat, tissue thermal conductivity, heat And the transfer coefficient.

The approximate expression of the skin temperature recovery curve is as follows:
Furthermore, it is also possible to obtain an expression that approximates the skin temperature recovery curve, and it is also possible to use such an expression.
Basically, such an expression becomes a complicated expression due to an increase in the number of terms, and it is a complicated operation to obtain the parameters of the expression. Therefore, as compared with the prior art, the above formula (2) or formula (2) that can faithfully reflect the pathophysiological function
And formula (3) is preferably used.

Next, by fitting the time-varying temperature of each pixel (i, j) obtained from the obtained time-series image data to the above equation (2), the temperature of each pixel (i, j) is obtained. Parameter value, α (i, j), μ1 (i,
j), β (i, j) and μ2 (i, j) are calculated. In this case, a fitting method using a nonlinear least squares method is used. In addition, as such a fitting method, Ga
uss-Newton method, Damping Gauss-Newton method, modified Marquard
There are t method, Simplex method and the like. Note that time-series image data may be fitted by substituting equation (3) into equation (2). In this case, the above-described fitting method can be used.

Further, by substituting equation (3) into equation (2),
The following equation is obtained.

(Equation 3) The equation (4) obtained in this way may be used instead of the equation (2).

Then, the above-described operation for obtaining each parameter value is performed for all the pixels by using the above equation (2) or (4), and the calculated parameter values are converted into different colors (different shades) for each pixel. (Including colors) shown in FIG. 2 (α, μ1, β, μ2). Here, calculation images (α, μ1, β,
μ2) is an α image (α), a μ1 image (μ1), a β image (β), and a μ2 image (μ2). In FIG. 2, the subject is shown as an image composed of pixels in n rows and m columns, but in actuality, a color corresponding to the parameter value (a color (shading different depending on the parameter value) )) Is included.

The α image (α) thus obtained, μ1
The image (μ1), β image (β), and μ2 image (μ2) are obtained by calculating the sum of a plurality of exponential functions as a detailed theoretical expression of skin temperature recovery characteristics after cooling load when obtaining these images. Since (2) is used, it can be said that the image is a pathophysiological function image from which the contribution of unnecessary factors to the disease state has been removed. Therefore, the accuracy rate of thermographic diagnosis using the α image (α), μ1 image (μ1), β image (β), and μ2 image (μ2) can be improved.

When the above equation (2) or the equations (2) and (3), which are the sum of a plurality of exponential functions, is used as an approximate equation of the skin temperature recovery curve, immediately after the end of the cooling load, A sufficiently satisfactory approximation to the data is possible,
The accuracy rate can be improved. Further, when the equation (4) in which the equation (3) is substituted into the equation (2) is used, it is not necessary to take an image until a steady state is reached, and only a plurality of thermal images are taken from t = 0. So in actual clinical applications,
It is more efficient to use this formula. That is, t = 0
Since a plurality of thermal images can be photographed from and the calculated image can be obtained from these thermal images, the calculated image can be obtained very easily.

Further, as described above, a plurality of parameter values are obtained, and a plurality of calculated images (α image (α),
μ1 image (μ1), β image (β), μ2 image (μ2)), and based on these calculated images (α, μ1, β, μ2), Calculation image (α, μ
1, β, μ2), it is possible to make a comprehensive judgment, and the accuracy rate can be improved.

In recent years, it has been revealed that pathological conditions such as growth of various solid tumors and metastatic tumors involve continuous angiogenesis. Thus, even in early cancers such as non-palpable breast cancer, certain lesions with angiogenesis
According to the present invention, a plurality of parameter images (α image (α), μ
One image (μ1), β image (β), μ2 image (μ2)) can be expected to be extracted as an abnormal pattern, and the result can contribute to an improvement in the accuracy of thermography diagnosis. That is, among the parameters, there are factors related to blood vessels such as blood flow as relevant factors. Angiogenesis affects parameter values, and an image based on angiogenesis may appear in a parameter image.

[0022]

According to the method for creating a calculated image in load thermography according to the first aspect of the present invention, based on a detailed theoretical formula of skin temperature recovery characteristics after cooling load, a time-series thermal image after loading is completed. The created calculation image of a plurality of parameters is a pathophysiological function image excluding the contribution of unnecessary factors to the disease state. Therefore, by using such a calculated image, it is possible to improve the accuracy rate in thermography diagnosis.

In recent years, it has become clear that pathological conditions such as development of various solid tumors and metastatic tumors involve continuous angiogenesis. Thus, even in early cancers such as non-palpable breast cancer, certain lesions with angiogenesis
The plurality of parameter images according to the present invention can be expected to be extracted as an abnormal pattern, and the result can contribute to an improvement in the accuracy of thermography diagnosis.

[Brief description of the drawings]

FIG. 1 is a drawing for explaining time-series thermal image data used in a calculation image creation method in load thermography according to an example of an embodiment of the present invention.

FIG. 2 is a diagram for explaining calculation images (α image, μ1 image, β image, μ2 image) of parameters obtained by a calculation image creation method in the load thermography of the above example.

[Explanation of symbols]

 α α image (calculated image) β β image (calculated image) μ1 μ1 image (calculated image) μ2 μ2 image (calculated image) to-tn Time-series thermal image (time-series temperature distribution image)

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-52-94697 (JP, A) JP-A-4-66822 (JP, A) JP-A-2-289238 (JP, A) JP-A 62-946 117527 (JP, A) Journal of the Japanese Society of Medical Radiology Physics, Vol. 10, No. 3, pp. 101-106 (58) Fields investigated (Int. Cl. ⁷ , DB name) A61B 5/00

Claims (1)

(57) [Claims] 1. Subjects after applying loads such as cooling and heating
Showing the temperature distribution on the skin surface and its temporal change,
A time-series temperature distribution image consisting of pixels in rows and m columns and the image after loading
Create a calculated image from an expression that approximates skin temperature recovery characteristics
In the calculation image creation method, the temperature is changed with the passage of time for each pixel determined from the time-series temperature distribution image, and the skin temperature recovery characteristics after loading is approximated by the sum of a plurality of exponential functions. Calculating a plurality of parameter values for each pixel, and then generating a calculation image indicating the parameter values for each pixel by color difference for each of the plurality of parameters. .