HYPERSPECTRAL IMAGE VISUALIZATION IN PATIENTS WITH MEDICAL CONDITIONS

Abstract

The present application discloses a novel algorithm to convert medical hyperspectral images (MHSI) into RGB (RedBlueGreen) images in different medical conditions by making use of the three spectral bands (Red, Green and Blue) of the MHSI and mapping them into Red, Green and Blue components for visualization of hyperspectral images.

Claims

1. A method, comprising selecting a biological tissue area for investigation; scanning under a medical hyperspectral imaging device; collecting a medical hyperspectral image (MHSI) of the biological tissue area under investigation; converting a MHSI to a color image as a RGB image of the biological tissue area under investigation; and visualize the RGB image to characterize the biological tissue under investigation.

2. The method of claim 1, wherein a MHSI is converted to a RGB image by a scaling transformation.

3. The method of claim 2, wherein the scaling transformation is via a gamma correction factor.

4. The method of claim 2, wherein the scaling transformation is via a linear mapping.

5. The method of claim 1, wherein a MHSI is converted to a RGB image by a XYZ to RGB matrix transformation.

6. The method of claim 1, wherein the biological tissue area for investigation is diabetic foot ulcers in patients with diabetes.

7. The method f claim 1, wherein the biological tissue area for investigation is tumor.

8. The method of claim 5, wherein the XYZ to RGB matrix transformation helps in mapping the regions of oxy-hemoglobin and deoxy-hemoglobin to characterize the biological tissue under investigation.

9. The method of claim 8, wherein the XYZ to RGB matrix transformation utilizes the spectrum from MHSI to determine accurate oxygenation level, with an error of less than 0.1% to 0.4%.

10. A method of claim 1, further comprises: converting MHSI to RGB images for visualization of foot ulcers in patients with diabetes.

11. A method, comprising selecting a biological tissue area for investigation; collecting a medical hyperspectral image (MHSI) of the biological tissue area under investigation; converting a MHSI to a color image as a RGB image of the biological tissue area under investigation by a XYZ to RGB matrix transformation; and visualize the RGB image as produced.

12. A process, comprising: collecting the MHSI of the biological tissue from a patient suffering from a medical condition; predicting the three hyperspectral bands of the visible spectrum; calculating the displayed luminance for each pixel value by applying a scaling transformation to all the three selected bands; and producing a sub-spectral color image by using XYG to RGB matrix transformation.

13. The process of claim 12, wherein the scaling transformation is via a gamma correction factor.

14. The process of claim 12, wherein the scaling transformation is via a linear mapping.

15. The process of claim 12, wherein the three hyperspectral bands are red, green and blue bands of the visible spectrum.

16. A method, comprising: selecting the biological tissue or area of biological tissue to be diagnosed; collecting information for vital signs and other parameters relevant to a particular medical condition; generating a MHSI; converting the MHSI to a RGB image; and identifying areas of biological tissue that are at a risk for a medical condition.

17. The method of claim 16, wherein the biological tissue area to be diagnosed is diabetic foot ulcers in patients with diabetes.

18. The method of claim 16, wherein the biological tissue area to be diagnosed is tumor.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Example embodiments are illustrated by way of example and no limitation in the tables and in the accompanying figures, like references indicate similar elements and in which:

[0024] FIG. 1 shows the MHSI formation and converting MHSI to RGB image.

[0025] FIG. 2 shows a visualization model for MHSI to RGB image transformation.

[0026] FIG. 3 shows a flow-chart representation of steps for converting MHSI into RGB images.

[0027] FIG. 4a, 4b, 4c, 4d shows the three MHSI slices of Blue (4a), Green (4b) and Red (4c) bands respectively selected from MHSI cube of images and the output image (4d) produced.

[0028] FIGS. 5A and 5B shows the comparison study between MHSI and RGB images in plantar and dorsum regions of the foot of a patient with diabetic foot ulcers respectively.

[0029] Other features of the present embodiments will be apparent from the accompanying figures and from the detailed description that follows.

DETAILED DESCRIPTION

[0030] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the art to which the invention belongs.

[0031] The term ‘gamma correction factor’ as used herein is one of the non-linear operation which is used to code and decode luminance in videos and still image system. It helps in maximizing the use of the bits or bandwidth relative to how humans perceive light and color (3) and thus components for properties of human vision.

[0032] The term ‘vital signs’ as used herein corresponds to measures of different physiological statistics of a patients/person to assess the basic body functions such as body temperature, pulse rate, heart rate, blood pressure, respiratory rate and other basic parameters depending on the medical condition.

[0033] The term ‘linear mapping’ as used herein is a function between 2 modules that preserves the operations of module addition and scalar multiplication.

[0034] The term ‘RGB image’ as used herein is a color model comprising of components of Red, Green and Blue (RGB) light from the visible region of electromagnetic spectrum. The resulting image as produced is a combination of the three primary RGB colors with different degrees of lightness (4).

[0035] The term ‘scaling transformation’ as used herein is a transformation which scales the coordinates of an object. It is specified either by working directly with the local coordinates or by expressing the coordinates in terms of frames (5).

[0036] The term ‘medical condition’ is used in its scientifically accepted sense to refer a disease condition as compared to a normal condition. The ‘medical condition’ as used herein may include Cancer, Tumors, Diabetes Foot Ulcers, Ulcers, Skin diseases, Neurological diseases, Vascular diseases, Muscular diseases, Cardiovascular diseases among other known disease conditions.

[0037] Although the present embodiments have been described with reference to specific example embodiments, it will be evident that various modifications and changes may be made to these embodiments without departing from the broader spirit and scope of the various embodiments.

[0038] Several embodiments of the present disclosure discuss a unique algorithmic way to convert MHSI of a biological tissue to a RGB image for a better visualization.

[0039] HSI spectroscopy since its discovery finds many applications ranging from agriculture, astronomy, chemical, environment and now finds its use in medical imaging. HSI was originally defined as a spectral sensing technique which acquires hundreds of contiguous narrow waveband images in the visible and IR regions of electromagnetic spectrum (6). The use of the technique has further extended into physiology and pathology areas of medicine to understand changes in living tissues in animal and human studies and thus is known as MHSI. As HSI depicts changes in material composition or gives information on different material present in the environment similarly MHSI gives information on chemical composition of biological tissues. The primarily absorbers are oxy- and deoxy-hemoglobin and thus MHSI characterize the in-vivo absorption spectra of these compounds in biological tissue by differentiating between the light absorbed and reflected by oxy-hemoglobin and deoxy-hemoglobin. The technology provides very important information regarding different physiological parameters such as oxygen delivery, oxygen extraction, total hemoglobin and water with spatial patterns at the level of microcirculation. Thus, the spectral and spatial features of MHSI can be combined to differentiate between two different states of biological tissue usually a normal tissue and a diseased tissue. However, the complexity of the results and the cost associated with MHSI as a whole is a disadvantage to the analytical technique. However, transforming the images from MHSI to RGB images will help in evaluating the complex MHSI into a more user friendly, easily evaluated RGB images. Further, it will help in making clinical decisions by monitoring the diseased and normal tissue and designing the mode of therapy.

[0040] Various embodiments of the invention discuss the use of disclosed algorithm to convert MHSI to RGB images which are user friendly, monitor and computer friendly, less complicated and more informative for predicting oxygen levels in a biological tissue of patients with medical conditions.

[0041] FIG. 1 shows the formation of MHSI and further conversion of MHSI to RGB images. Hyperspectral imaging device 102 works in the visible and infrared (IR) regions of the electro magnetic (EM) spectrum 101 and thus are able to produce 2D MHSI 105. MHSI is a collection or a set of different images each corresponding to a particular wavelength band. It is based on the concept that each pixel from the biological tissue under investigation contains detailed information about the spectrum of the incoming light. It acquires the image in cubes of data comprising rich information regarding the biological tissue in the image. Thus the pixels collected from the sample is in the form of a 3-D hyperspectral data cube 103 which is then converted into 2D MHSI 104. 2D MHSI are in a mix of colors which result in increase in its complexity for analysis and may result in poor analysis of the image obtained especially in case of human tissue analysis may affect the treatment decisions. Thus, to make the MHSI more user friendly and to simplify the image for a better interpretation, is converted into RGB image 105 with the help of disclosed algorithm. The developed algorithm utilizes the MHSI spectrum to determine accurate oxygenation level, with an error of less than 0.2%.

[0042] FIG. 2 shows the visualization model to convert MHSI into RGB images. A particular wavelength (λ) from all the three bands such as red (i,j, 201), green (i,j, 202) and blue (i,j, 203) is selected which is from the visible spectrum of the image. The values are then processed in the visualization model 204 to convert the MHSI to a RGB image 205. To convert the images for visualization, the RGB matrix produced from MHSI (RGB i,j) is subjected to an algorithm as disclosed in the present application. The visualization model 204 as shown is FIG. 2 comprise of predicting the luminance for each pixel value as selected and applying either the gamma correction factor to the 3 hyper spectral bands RGB (i,j) or by applying linear mapping to the 3 hyperspectral bands RGB (i,j) to produce 1 single color image depicting red, green and blue components of the visible spectrum.

[0043] To produce RGB image from MHSI, a scaling transformation via linear mapping is used to scale the HS matrix elements to displayable values. For any MHSI matrix, M, the linear mapping function applies a linear transformation to its elements, with the lower value mapping to 0.0 and the highest value mapping to 1.0 to 2.0. The linear transformation as disclosed in the present application is given by the equation:

[00001]$\begin{array}{cc}{M}_{\mathrm{lmap}}\left(i,j\right)=\frac{M\left(i,j\right)-M_{\min }}{M_{\max }-M_{\min }}& \mathrm{Equation}.Math..Math.1\end{array}$

[0044] Following, a sub-spectral resolution of color images is produce from MHSI by using the XYZ to RGB matrix [RGB (i,j)] transform as disclosed:

sRGB(i, j)=XYZ_sRGB(M.sub.lmap(i, j, λ))   Equation 2

[0045] FIG. 3 shows the steps involved to carry out the disclosed invention. For this, a user or a trained technician may start the imaging apparatus 301 which may be any hyperspectral imaging apparatus available commercially or non-commercially. A user may then select the biological tissue area to be investigated in patients with a medical condition 302. The area of a biological tissue to be investigated is then exposed to the imaging device to collect the MHSI 303. From the 2D MHSI (303) select the Red, Green and Blue bands 304 and predict the matrix for each band as Red (i,j), Green (i,j) and Blue (i,j) respectively 305. Calculate the luminance to be displayed for each pixel value from the MHSI coordinates by either applying the gamma correction factor 306 or by applying the linear mapping to the band matrix 307. The values so obtained are put into the disclosed algorithm of XYZ to RGB matrix transformation 308 to produce a sub-spectral resolution of color images. Collect the RGB images converted from the MHSI 309. A user/physician may then analyze the RGB image 310 to determine the accurate oxygenation and de-oxygenation saturation levels in the biological tissue under investigation. The information may then help in analyzing the condition of the tissue, extent of tissue damage, onset of ulcers or tumors, degree of tumor or ulcers growth. A user may also extract information such as oxygen delivery, oxygen extraction, total hemoglobin, water among other information. The information obtained can then be combined with vital signs, pressure readings and other relevant information to get a clear and general picture of patient's medical condition.

[0046] FIGS. 4a, 4b, 4c and 4d show the use of disclosed algorithm to convert the MHSI images produced from the known technology of University of Nottingham to RGB image in condition of foot ulcers in patients with diabetes. The figure shows different slices from MHSI corresponding to 3 different bands of visible region. Blue (4a), Green (4b) and Red (4c) bands are selected from the 3D-MHSI and the disclosed transformation algorithm (equation 2) is applied to the band matrix to derive a RGB image (4d). The image (4d) as shown in FIG. 4 helps in predicting the saturation levels of oxygenation and de-oxygenation as they can differentiate the levels with simple color contrast rather than complex color analysis.

[0047] FIGS. 5a and 5b shows the comparison studies of RGB images and MHSI of foot ulcers from patients with diabetes. Ulcers in plantar (5a) and dorsum (5b) regions of the foot of the diabetic patient are being investigated. The figure demonstrate HSI produced by University of Nottingham using algorithm developed by them through Matlab using Mathworks® and RGB images produced by inventors using the disclosed algorithm. The images can determine the accurate oxygenation and de-oxygenation levels in the plantar and dorsum regions of the foot ulcers in patients with diabetes with easy visualization.

[0048] While the present disclosure has been described with reference to an exemplary embodiment, changes may be made within the purview of the appended claims, without departing from the scope and spirit of the present disclosure in its aspects. Also, although the present disclosure has been described herein with reference to particular materials and embodiments, the present disclosure is not intended to be limited to the particulars disclosed herein; rather, the present disclosure extends to all functionally equivalent structures, methods and uses, such as are within the scope of the instant claims. Accordingly, the specification and drawings are to be regarded in an illustrative rather than in a restrictive sense.