Image analysis – Applications – Biomedical applications
Reexamination Certificate
1998-08-20
2001-04-10
Mehta, Bhavesh (Department: 2721)
Image analysis
Applications
Biomedical applications
C382S165000, C382S164000
Reexamination Certificate
active
06215893
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to medical equipment and methods, and more particularly to an apparatus and a method for measurement and temporal comparison of skin surface images.
BACKGROUND OF THE INVENTION
The current rise in skin cancer incidence and the increased public awareness of the dangers of skin cancer has reinforced a need for tracking of skin lesions in a reliable and accurate fashion. Most important is the need for a method of early detection of melanoma, the deadliest skin cancer. Until recently, when a physician noted a lesion on a patient the method of recording was through painstaking notes and measurements. However, this is not necessarily an objective nor sufficiently stringent method of follow-up and may be insufficient for accurate diagnosis.
In order to standardize the notetaking procedure, the clinical ABCD system for checking static clinical features was introduced by the NYU Medical Center. In this system a lesion is checked for Asymmetry, Border irregularity, Color variegation and Diameter. Diagnosis based on the clinical ABCD system has achieved a sensitivity of roughly 80%. This low sensitivity has led to the practice of removing almost any atypical mole. In addition, the ABCD system can only be applied to moles greater than 6 mm in diameter. It would be desirable to detect malignant moles at an even earlier stage.
In Great Britain, the Seven Point Checklist was introduced to improve the sensitivity of the clinical exam. In this method, three major features (change in size, shape or color) and four minor features (inflammation, crusting or bleeding, sensory change and diameter greater than 7 mm) are assessed. The inclusion of both static and dynamic change has led to a higher sensitivity in diagnosis. The emphasis on change in this method shows the dynamics of the lesion. However, the system has not gained wide acceptance because there have not been enough studies in order to quantify what rate of changes are considered alarming, since there are various rates of change which are acceptable and these rates may vary according to the individual. In addition, applications of this system have led to low specificity. Some of the parameters of the checklist are subjective, such as the sensory ones, and hence may not be repeatable, thus causing a problem in developing an objective diagnosis/monitoring system.
Photography has been used to aid in record keeping, but subtle changes in angle, lighting, skin tension and distance can affect the reading and therefore cause misdiagnosis. In addition, photographs, especially polaroids which are commonly used, can show color degradation over time. The newest methods of skin image tracking rely on epiluminescence microscopy (ELM). This technology looks deeper into the skin and is able to image structures up to the dermis-epidermis junction. In depth investigations have led to deduction of relevant features bearing a high correlation with malignancy. As the technique allows visualization of features not visible to the naked eye, using this technique results in higher sensitivity and specificity than that obtained with the clinical ABCD system. Typical values obtained with this method range from 92-100% for the sensitivity and 80-92% for specificity.
ELM, although better than previous techniques, has not yet been brought to the necessary level of sensitivity and specificity. There is a lack of exact correlation with structures seen in the histology. In current applications of epiluminescence microscopy technique the image capturing process involves pressing a transparent glass or plastic on the skin. This can introduce error into the system since different applications of pressure cause a different stretching of the skin which is imaged onto the camera. As there is no regulation of the pressure, the images captured are not fully repeatable. In addition, the other features by which the current ELM technology operates are extremely difficult to obtain in an automated way in a repeatable fashion. Various features which are important in the ELM are dynamic by nature and take some time before they develop. Therefore, findings may be misinterpreted due to lack of exact staging of the mole or external factors affecting the interpretation. Thus the interpretation of the features is somewhat difficult. In some systems, the lesion area is determined by manually selecting an appropriate threshold to determine the skin/lesion border. As all the parameters are then measured automatically based on this choice, the results may be biased and therefore not repeatable. Therefore, it is necessary to provide a robust system of parameters for use in diagnosis.
Other areas in which temporal comparison of skin surface images may be of use include cosmetics, skin treatment and identification. To this end a method of storing an image which is repeatable and comparable would be desirable.
Thus, it would be desirable to provide a fully automatic, sensitive, specific and cost-effective system for measurement and temporal tracking of lesion skin images which would be repeatable and comparable.
SUMMARY OF THE INVENTION
Accordingly, it is a principal object of the present invention to overcome the disadvantages associated with current ELM and clinical technology and to provide a method for temporal comparison of skin surface images which is fully automatic and based on robust parameters.
In accordance with a preferred embodiment of the present invention, there is provided an apparatus for temporal comparison of skin surface images comprising:
means for obtaining a digital image of a skin area and storing same for analysis and display;
means for determining in iterative fashion a region of interest within said digital image according to a set of predetermined image color parameters;
means for separating said region of interest into data values representing color clusters and color islands;
means for mapping said data values to be associated with respective skin and lesion data sets;
means for extracting from said respective data sets measurements corresponding to a set of predetermined image measurement parameters;
means for performing statistical calculations on said respective extracted data set measurements to obtain a set of numerical factors associated with predetermined image analysis categories;
means for storing said set of numerical factors;
means for analyzing said set of numerical factors for diagnosis determination; and
means for displaying said digital image, said numerical factors and said diagnosis determination.
In a preferred embodiment of the invention, a CCD camera is provided with telecentric optics enabling the capture of images with constant magnification independent of the camera-object distance. Therefore, the sensitivity to height location of the camera above the object is markedly reduced. The measurement system is non-contact and non-destructive. The camera is linked to a PC with a frame grabber for storing the image and processing means for calibrating light and for calculating statistics of the image. The illumination source provides a homogeneous illumination of the skin. The light is calibrated by eliminating the effects of the illuminating light from the assumed white, spatially invariant light, and then correcting the acquired image of the lesion. A model has been provided which transforms an acquired image (AI) of an object into the model image (MI) which would have resulted if illumination of the object were solely by the reference (white) light.
Three monochromatic images are produced, one for each of three color channels. The three images are grabbed and processed, and a combination of these three images gives the true color (red, green and blue or RGB) image. The RGB image is then transformed to the CIE-Lab representation. An algorithm uses the image and reference image to calculate the skin/lesion border which the eye would suggest if the lesion was illuminated by a standard white light. It separates the image into two parts, the largest lesion and the rest of the skin. The image is proces
Kutscher Tuvia
Leshem Ze'ev
Langer, Pat. Atty. Edward
Mehta Bhavesh
Romedix Ltd.
LandOfFree
Apparatus and method for measurement and temporal comparison... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for measurement and temporal comparison..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for measurement and temporal comparison... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2552823