ES2577011A1 - Procedure for reading the tuberculin reaction by thermal infrared imaging (Machine-translation by Google Translate, not legally binding) - Google Patents

Abstract

Procedure for reading the tuberculin reaction by infrared thermal imaging for the measurement of the size of a skin induration comprising the following steps: - injection of purified tuberculin protein derivative (ppd) or related substance in a subject, - taking the first infrared image of the skin, - conversion of the first image taken in a scaled two-dimensional array to a second image, - reference object detection, - obtaining central pixel and radius of the object, - calculation of conversion of pixels to millimeters, - i reject internal pixels of the object, - intensity transformation of the pixels of the second image, obtaining a third image, - segmentation of a third image in three regions according to local growth method, - area calculation, maximum, minimum and standard deviation of regions and depth. (Machine-translation by Google Translate, not legally binding)

Description

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

Procedure for reading the tuberculin reaction by infrared thermal imaging.

OBJECT OF THE INVENTION

The present invention relates to a method for reading the tuberculin reaction by infrared thermal imaging for the measurement of the size of a skin induration.

BACKGROUND OF THE INVENTION

According to data provided by the World Health Organization in 2004, two billion people are infected by the bacillus of tuberculosis (TB). It is estimated that 8 million new cases appear every year.

TB is a contagious disease caused by a bacillus and without absolute control, with large differences between poor and rich countries due, among other causes, to hygienic, food and sanitary differences. The diagnosis of tuberculosis is based on the interpretation of the cutaneous immunological reaction. Thus, infection with Mycobacterium tuberculosis provides a delayed immune response that is measured with the American Thoracic Society and Centers for Disease Control And Prevention: Diagnostic standars and classification of tuberculosis in adults and children. Crit Care Med 2000 161: 1376). The reaction consists of an area of small induration (elevation of the surface of the skin with respect to the adjoining area that can be visualized and palpated by a simple scan) that appears at the site of the skin test. The size of the induration measured in millimeters indicates whether the test is negative or positive.

Detection methods such as palpation or marking have been used (Bouros D, Zeros G, Panaretos C, Vassilatos C, Siafakas N. Palpation vs pen method for the measurement of skin tuberculin reaction (Mantoux test). Chest. 1991 99: 416 -9.) But the interpretation is subjective (Pouchot J, Grosland A, Collet C. Coste J, Esdaile JM, Vinceneux P. Reliability of tuberculin skin test measurement. Ann Intern Med 1997 126: 210-4.).

On the other hand, there are precedents in the application of infrared thermal imaging as a control and diagnostic measure in medicine in fields other than the present invention (Bagavathiappan S, Sarananan T, Philip J, Jayakumar T Ph J, Baldev R , Karunanithi R, Panicker TMR, Korath MP, Jagadeesan K. Infrared thermal imaging for detection of peripheral vascular disorders. J Med Phys 2009 34: 43-47.). This technique measures radiated electromagnetic energy between the wavelength range 3-12 pm. An increase in infrared radiation is caused by high blood flow, increased metabolic activity and various inflammatory reactions that take place in the affected area (Schaefer AL, Cool N, Tessaro SV, Deregt D, Desroches G, Dubeski PL, Tong AKW , Godson DL Early detection and prediction of infection using infrared thermography. Can J Anim Sci 2004 84: 73-80.).

In the inflammatory process, the temperature increase, which can be measured by infrared, is caused by the increase in blood flow with vascular dilatation, blood proteins and cellular extravasation (Ring FE, Ammer K, Thermal imaging in diseases of the skeletal and neuromuscular systems.In Medical infrared imaging.Edited by NA Dinkides, JD Bronzino CRC, US. 2008: 17-1.).

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

In general, the reaction after the injection of purified tuberculin protein (PPD) is characterized by edema, leukocyte infiltration and lymphocyte and mononuclear cells as well! as extravasation of erythrocytes in the epidermal and dermal layer of the skin (12. Haholu A, Ciftci F, Karabudak O, Kutlu A, Bozkurt B, Baloglu H. The significance of histopathologic patterns in positive tuberculin skin test. Journal of Cutaneous Pathology. 2008 35: 462-465.).

DESCRIPTION OF THE INVENTION

The present invention is established and characterized in the independent claims, while the dependent claims describe other features thereof.

The object of the invention is a procedure for reading the tuberculin reaction by infrared thermal imaging that measures the size of a skin induration. The technical problem to solve is to configure the steps of the procedure to achieve the mentioned object.

Image processing techniques are used, such as the Canny method cited in the claims referring to the well-known Canny algorithm for the detection of object edges that uses a filter based on the first derivative of a Gaussian to eliminate noise from an image . Also, the Hough circular transform is a known technique for detecting figures in digital images.

An advantage of the procedure is that it can be implemented in any device such as a mobile phone, with the advantages of portability, which makes it useful in isolated areas such as populated with infected subjects. This ease of implementation is due to the fact that the calculation needs to carry out the procedure of this invention can be carried out by means of the processors of a mid-range mobile phone at the time of the presentation of the invention and to the fact that the price of the Type of camera needed for its realization is low.

DESCRIPTION OF THE FIGURES

The present specification is complemented, with a set of figures, illustrative of the preferred example and never limiting the invention.

Figure 1 represents a front view of a forearm of a subject with a circular reference object, a quadrangular analysis area mask and an induration with two distinct regions.

DETAILED EXHIBITION OF THE INVENTION

An embodiment of the invention is shown below with support in the figure.

The Mantoux tuberculin cutaneous test (TST) is carried out by intradermal injection of 0.1 ml of purified tuberculin protein derivative (PPD) into the skin (1) of the subject, for example in this embodiment in the forearm. The TST is read 48 hours after the injection, by inspection, palpation and induration measurement. An induration size greater than 1 cm is considered a positive result, indicating the presence of Mycobacterium tuberculosis infection.

The procedure prior to the reading of the TST by infrared thermal imaging is as follows. The subjects remain in the laboratory for at least 15 minutes to reach a thermal equilibrium within the laboratory's room. The temperature, humidity and circulation of the air in the laboratory are maintained around 23 ° C ± 1 and a degree of humidity

3

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

around 50%, without roof air filter. Subjects are asked not to consume hot drinks or food at least one hour before the session, and not to use cutaneous preparations such as creams or talcum powder. The emissivity of the skin (1) is assumed at 0.98 ± 0.01. The lighting is maintained with wavelengths no greater than 1pm.

In this embodiment, thermal infrared images are obtained from the front face of each subject's forearm. You can use any infrared camera, such as a TiR32 Fluke® or even a camera included in a mobile phone (after removing the filters that include these cameras, whose purpose is precisely to eliminate these frequencies). The distance from the camera to the skin (1), the forearm in this case, is 40cm. The reference object (2) used in this embodiment is a 24mm diameter coin. The acquisition and saving of images is done with the SmartView® software, which can be used with a mobile phone. Each image is exported in jpg format. The temperature measurements are exported to a text file. Matlab 2014b® is used as an image treatment program. Note that the algorithm can be reprogrammed in Java or C language, which are programming languages in which applications can be made on mobile devices.

The invention is a method for the detection of tuberculin by infrared thermal imaging for the measurement of the size of a skin induration comprising the following stages:

- 0.1 ml injection of purified tuberculin protein (PPD) derivative into a subject, in the exposed embodiment it is done in the subject's forearm although it could be another suitable area, such as a thigh, etc., also the PPD is the substance more commonly used and whose implementation is preferred in this case, although other substances related to PPD can be used, such as ESAT6 / CFP10 (Diaskintest®) M tuberculosis specific antigens,

- taking a first infrared image of the skin (1) of the subject,

- conversion of the first image taken in a two-dimensional matrix scaled according to: second image = (temp-Tmin) / (Tmax-Tmin), obtaining values between 0 and 1 of a second image in grayscale, with Tmin and Tmax being temperatures minimum and maximum respectively,

- detection of a reference object (2) in the second image, for example circular, applying edge detection, for example by the Canny method and the circular Hough transform,

- obtaining the central pixel and the pixel radius of the reference object (2),

- calculation of the conversion factor of pixels to millimeters by knowledge of the radius in millimeters of the reference object (2),

- rejection of the internal pixels of the reference object (2) by setting their values as 0 in the second image,

- intensity transformation of the pixels of the second image: assigning 0 when they are below 0.5, passing those from the range 0.5-1 to the range 0-1, obtaining a third image,

- segmentation of the third image into three regions: first (4), corresponding to induration, and second (5), corresponding to erythema, called as! Due to the similarity they have with the anatomo-pathological findings produced by the tuberculin reaction, mentioned above, the third corresponds to the background (6). A method of local growth of regions is used comprising the following stages:

- selection of an initial seed pixel with maximum temperature (PTmax) within a rectangular area centered on it,

- binarization of the third image, saving the pixels of the skin (1) of the subject, with value 1, rejecting the pixels outside the skin (1) of the subject, with value 0, obtaining a fourth binary image,

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

- creation of an analysis area mask (3) from the fourth binary image, centered on the maximum temperature seed pixel (PTmax),

- iterations on the third image for the detection of first and second sub-regions corresponding to the first (4) and second (5) regions respectively. Since there may be several reaction zones, the first (4) and second (5) regions may be formed by several subregions. In each iteration a new first and second subregion is detected from each new seed pixel. Each subregion is formed by adding neighboring pixels according to a criterion to the corresponding seed pixel for each region (4,5). The preferred option consists of: first, neighboring pixels are added to the seed pixel to grow the first subregion, as long as there are pixels that meet the following criteria:

-AxTh1 <minS1 (ij) <Th1 ^ -S1 (ij) = reg_mean1-temp_e (i, j),

min (| S2 (ij) |) <Th2 ^ -S2 (ij) = reg_mean2-temp_e (i, j),

where reg_mean1 is the average intensity value of all pixels that already belong to the first region (4), temp_e (i, j) the intensity value of a neighboring pixel, S1 (i, j) the similarity value of a pixel neighbor, Th1 the threshold used to delimit whether a neighboring pixel joins the first subregion or not, and the scaling parameter for the lower limit of the similarity value. When there are no longer neighboring pixels that meet the above criteria, a second criterion is applied to add pixels to the second subregion, defined as:

min (| S2 (ij) |) <Th2 ^ -S2 (ij) = reg_mean2-temp_e (i, j) where reg_mean2 is the average intensity value of all pixels that already belong to the second region (5), temp_e ( i, j) the intensity value of a neighboring pixel, S2 (i, j) the similarity value of a neighboring pixel and Th2 the threshold used to delimit whether or not a neighboring pixel joins the second subregion. When there are no neighboring pixels that meet the second criteria, another seed pixel is selected and a new iteration begins. A minimum size, for example 5 mm, is set for the first subregion of the first seed pixel. The first subregions of the rest of the seed pixels must have that same minimum size or be adjacent to the first subregion of the first seed pixel. Each new seed pixel must meet the similarity criteria of the first region (4) and be located at a maximum distance from the first subregion obtained in the first iteration,

- continuation of the iterations until a new seed pixel is not found,

- union of the adjacent first sub-regions,

- calculation of the maximum distances of the first subregions,

- calculation of the maximum transverse distances to the maximum distances of the first subregions,

- elimination of the first sub-regions whose relationship between their maximum transverse distance and their maximum distance (d) does not exceed a threshold, for example of 0.35,

- calculation of the maximum distance (d) of the first region (4) as the sum of the maximum distances of the first subregions,

- calculation of the maximum distance (D) of the second region (5),

- calculation of the background (6) eliminating the pixels that belong to the first (4) and second (5) regions according to the analysis area mask (3),

- calculation of the area, the maximum temperature, the minimum, the mean and the standard deviation of the three regions (4,5,6),

- union of the second region (5) with the bottom (6) if the following condition is met: tmean2-tstd2-tmean3-tstd3 <= 0.3, with tmeank and tstdk (k = 2 or 3) the average temperatures and deviations standard of the second region (5) and the bottom (6), respectively,

- union of the first region (4) with the second (5) and the bottom (6) if, in addition to fulfilling the condition of the previous point, the following condition is met: tmean1-tstd1-tmean23-

tstd23 <= 0.25 and tmean1-tstd1-tmean2-tstd2 <= 0.3 or tmean2-tmean3> = 0, with tmeank and tstdk (k = 1, 23, 2 or 3) the average temperature and the standard deviation of the first regions ( 5), second (5) with bottom (6), second (5), and bottom (6), respectively.

5 Optionally, the analysis area centered on the maximum temperature seed pixel (PTmax) has dimensions of 5cm x 5cm.

Another option is that both Th1 and Th2 are normalized by dividing them by the maximum and minimum temperature difference of each subject. In addition, Th1 and A can take three different values: low, for example 0.49, medium, for example 0.6, and high, for example 0.71 based on

The needs of each case.

Another option is that the minimum size of the first subregion of the first seed pixel is 5mm.

15 Another option is that the maximum distance from the first subregion of the first iteration (4)

It is 5mm.

Claims (11)

5 10 fifteen twenty 25 30 35 40 Four. Five fifty 1. -Procedure for reading the tuberculin reaction by infrared thermal imaging for the measurement of the size of a skin induration characterized by comprising the following stages: - injection of 0.1 ml of purified tuberculin protelna derivative (PPD) or a substance related to said protelna into a subject, - taking a first infrared image of the skin (1) of the subject, - conversion of the first image taken in a two-dimensional matrix scaled according to: second image = (temp-Tmin) / (Tmax-Tmin), obtaining values between 0 and 1 of a second image in grayscale, with Tmin and Tmax being temperatures minimum and maximum respectively, - detection of a reference object (2) in the second image by applying edge detection, - obtaining the central pixel and the pixel radius of the reference object (2), - calculation of the conversion factor of pixels to millimeters by knowledge of the radius in millimeters of the reference object (2), - rejection of the internal pixels of the reference object (2) by setting their values as 0 in the second image, - intensity transformation of the pixels of the second image: assigning 0 when they are below 0.5, passing those from the range 0.5-1 to the range 0-1, obtaining a third image, - segmentation of the third image into three regions: first (4), second (5) and background (6), using a method of local growth of regions comprising the following stages: - selection of an initial seed pixel with maximum temperature (PTmax) within a rectangular area centered on it, - binarization of the third image, saving the pixels of the skin (1) of the subject, with value 1, rejecting the pixels outside the skin (1) of the subject, with value 0, obtaining a fourth binary image, - creation of an analysis area mask (3) from the fourth binary image, centered on the maximum temperature seed pixel (PTmax), - iterations on the third image for the detection of first and second sub-regions corresponding to the first (4) and second (5) regions respectively, of each new seed pixel, adding neighboring pixels, - continuation of the iterations until a new seed pixel is not found, - union of the adjacent first sub-regions, - calculation of the maximum distances of the first subregions, - calculation of the maximum transverse distances to the maximum distances of the first subregions, - elimination of the first sub-regions whose relationship between their maximum transverse distance and their maximum distance does not exceed a threshold, - calculation of the maximum distance (d) of the first region (4) as the sum of the maximum distances of the first subregions, - calculation of the maximum distance (D) of the second region (5), - calculation of the background (6) eliminating the pixels that belong to the first (4) and second (5) regions according to the analysis area mask (3), - calculation of the area, the maximum temperature, the minimum, the average and the standard deviation of the regions and the bottom. 2. - Procedure according to revindication 1 in which the formation of each subregion is formed by adding neighboring pixels according to a criterion to the corresponding seed pixel for each 5 10 fifteen twenty 25 30 35 40 Four. Five fifty region, consisting of: first, neighboring pixels are added to the seed pixel to grow the first subregion, as long as there are pixels that meet the following criteria: -AxTh1 <minS1 (ij) <Th1 ^ -S1 (ij) = reg_mean1-temp_e (i, j), min (| S2 (ij) |) <Th2 ^ -S2 (ij) = reg_mean2-temp_e (i, j), where reg_mean1 is the average intensity value of all pixels that already belong to the first region, temp_e (i, j) the intensity value of a neighboring pixel, S1 (i, j) the similarity value of a neighboring pixel, Th1 the threshold used to delimit whether a neighboring pixel joins the first subregion or not, and the scaling parameter for the lower limit of the similarity value. When there are no longer neighboring pixels that meet the above criteria, a second criterion is applied to add pixels to the second subregion, defined as: min (| S2 (ij) |) <Th2 ^ -S2 (ij) = reg_mean2-temp_e (i, j) where reg_mean2 is the average intensity value of all pixels that already belong to the second region, temp_e (i, j ) the intensity value of a neighboring pixel, S2 (i, j) the similarity value of a neighboring pixel and Th2 the threshold used to delimit whether or not a neighboring pixel joins the second subregion. When there are no neighboring pixels that meet the second criteria, another seed pixel is selected and a new iteration begins. A minimum size is set for the first subregion of the first seed pixel. The first subregions of the rest of the seed pixels must have that same minimum size or be adjacent to the first subregion of the first seed pixel. Each new seed pixel must meet the similarity criteria of the first region (4) and be located at a maximum distance from the first subregion obtained in the first iteration. 3. - Procedure according to claims 1 or 2 in which after the calculation of the area, the maximum temperature, the minimum, the average and the standard deviation of the regions and the bottom, the second region (5) joins the bottom (6) if met: tmean2-tstd2-tmean3-tstd3 <= 0.3, with tmeank and tstdk (k = 2 or 3) being the average temperatures and standard deviations of the second region (5) and the bottom (6), respectively. 4. - Procedure according to claim 3 in which the first region (4) is joined with the second (5) and the background (6) if it is met: tmean1-tstd1-tmean23-tstd23 <= 0.25 and tmean1-tstd1- tmean2- tstd2 <= 0.3 or tmean2-tmean3> = 0, where tmeank and tstdk (k = 1, 23, 2 or 3) are the average temperature and the standard deviation of the first (4), second (5) bottom regions (6), second (5), and background (6), respectively. 5. - Procedure according to claim 1 in which the detection of the reference object (2) is by the Canny edge detection method and the circular Hough transform. 6. Procedure according to claim 1 in which the analysis area centered on the maximum temperature seed pixel (PTmax) has dimensions of 5cm x 5cm. 7. Procedure according to claim 1 wherein the minimum size of the first subregion of the first seed pixel is 5mm. 8. - Procedure according to claim 1 in which the maximum distance from the first subregion of the first iteration (4) in 5mm. 9. - Procedure according to claim 1 in which the elimination of the first subregions is carried out when the relationship between its maximum transverse distance and its maximum distance does not exceed the threshold of 0.35. 10. Procedure according to claim 2 in which both Th1 and Th2 are normalized by dividing them by the maximum and minimum temperature difference of each subject. 11. -Procedure according to claim 2 in which Th1 and A take three different values: 0.49, 0.6 and 0.71.