A Preliminary Investigation on Ocular Thermal Image Patterns Using Clinical Case Studies
Dr. Sheeja V Francis
1, Dr. J.Sivakamasundari
1, Dr. Sandeep Jaipurkar
21.
Department of Biomedical Engineering, Jerusalem College of Engineering, Chennai-600 100.[email protected], 2. Vijaya Health Centre, Vadapalani, Chennai.
ABSTRACT: This paper presents a preliminary investigation on thermal patterns in normal and abnormal eyes using image processing techniques on ocular thermal images acquired from clinical cases. Initially, thermal images of face are acquired from normal and abnormal volunteers and the ocular Region of Interest (ROI) is segmented. Thermal contours are then extracted from the ROI and analyzed. It is observed that the temperature distribution and thermal contours exhibit a concentric pattern in normal eyes.
Whereas distinct deviation from this pattern is observed in the abnormal eyes. The inferences obtained from the study are promising and may be instrumental in developing a computer aided screening technique for early detection of eye abnormalities.
Keywords— Ocular Thermal Images; Ocular Thermal Patterns; Image Segmentation; Thermal Contours; Temperature Measurement
I. INTRODUCTION
The Information Technology revolution of the last decade and COVID-19 restrictions have driven communities online like never before. Education, health care, recreation, etc., are accessed through internet on computers and mobile phones. Prolonged exposure to illuminated displays of such electronic devices causes strain and fatigue on the eyes called Asthenopia[1]. The most common problem associated with eye fatigue due to aging and environmental factors is a condition called dry eyes [2].
These condition are often overlooked as simple tired eyes until it leads to severe vision problems.
Frequent eye checkup is always not feasible in terms of availability, time and cost. Hence there is an urgent need to develop low cost, non invasive screening system for detecting eye abnormalities. A normal eye has a stable tear film over the cornea and a uniform distribution in surface temperature. Whereas this film is unstable in abnormal eyes resulting in thermal variations [3]. As thermography can be used to detect variations in surface temperature [4] it has been used to detect Glaucoma[5] and diabetic eyes[6] by analyzing the ocular temperatures. This paper proposes to study the patterns of thermal variations in human eyes through clinical case studies. The aim of the case study is to test the feasibility of developing an automatic computer aided screening system for eye abnormalities from ocular thermograms.
A. Thermography
Infrared thermography has been widely used in fields such as information security, military research, surveillance and identity verification. It is a functional imaging procedure which senses the surface temperature distribution of an object by measuring the emitted infrared energy. Every object with
temperature above zero degrees emits infrared radiation. This energy is captured by a thermal camera which translates the detected incident infrared radiation into temperature readings. Actually the temperature difference between the object and its environment is calculated and visualized as heat maps called thermal images or thermograms.[7] The visualized thermal pattern of the object remains invariant to illumination changes in the environment. Objects that emit heat can still be detected even in total darkness making thermography best suited for night vision and surveillance applications.
As the thermal cameras generally use a picture sensor that features a different wavelengths of infrared, the images captured possess a single-color channel. These monochromatic images are also displayed in pseudo-color using a technique called density slicing. This is useful because although humans have much greater dynamic range in intensity detection than color, the ability to see fine intensity differences in bright areas is fairly limited [8]. Inbuilt camera software aids in processing and visualizing these fine differences using pseudo colouring. This presents the viewer with an approximate interpretation of the thermal profile of the object.
Over the last few decades the resoultion and accuracy of thermal cameras have been greatly improved through state of art technology. This has led to several investigations on medical applications of Thermography. The normal human body emits Infrared energy in the wavelength 8 – 14 µm [9] . This energy can be captured and displayed as normal heat patterns using a thermal camera. Conditions like fever, inflammation and pain manifest in an increase in body temperature and hence distinct deviations are observed in the thermal patterns [10]. Chronic conditions like cancer[11] and arthritis also show elevation in surface temperatures [12]. Hence over the last few decades, Infrared thermography is being widely explored as a promising imaging modality in the field of medical diagnostics.
Despite of much progress in detection of eye abnormalities for visible images, there has been little progress for thermal eye images. This work attempts to explore thermal patterns in normal and abnormal eyes using image processing algorithms on ocular thermograms.
II. MATERIALS AND METHODS
Thermal eye images of the 25 volunteers were acquired after obtaining informed consent atVijaya Health Centre, Vadapalani, Chennai by using thermal camera (ICI720P). The subjects include 19 volunteers aging between 10 – 55 years with no known eye defects . Images were also acquired from 6 abnormal volunteers with different ocular conditions such as cataract eye, diabetic eye, lens fitted eye and person with insomnia (lack of sleep). Also the thermal effects of prolonged mobile usage on the eyes was also studied. The details of patient data are presented in tables 1 and 2.
A. PATIENT DATA:
Table. 1 Normal Patient Data
Patient ID Age group (both male and female)
N001 10
N002-N007 19-21
N008-N013 23-25
N009-N019 35-55
Table. 2 Abnormal Patient Data Volunteers
ID Age Sex Abnormal
AN001 20 Male Eye sight
(power-3.57) AN002 67 Female Cataract eye
AN003 20 Male Insomniac eye
AN004 53 Male Diabetic eye
AN005 48 Male Presbyopia
TEST006 20 Male
Temperature difference for using mobile
phone
B. THERMAL CAMERA:
A thermal camera of Model: ICI720p with inbuilt software IR flash has been used to acquire the images from volunteers. All objects emit infrared radiation of varying intensities which is manifested as changes in heat and temperature. Thermal cameras detect changes in temperature by recognizing and capturing different levels of infrared light emitted by the objects. In a thermogram, the brighter colors such as red, orange, and yellow are used to indicate warmer temperatures while the purples and dark blue/black indicate cooler temperatures (less heat and infrared radiation emitted) [13].
Matlab software was used to analyse the ocular thermograms.
C. METHODOLOGY
The volunteers were prepared for image acquisition by following a standard protocol. They were asked to keep their face on a chin rest with eyes closed for 15 minutes ensuring a stable tear film and symmetric distribution of eye temperature. The temperature and humidity of the image acquisition suite
were controlled by an air conditioning unit at 25 degree centigrade. The images acquired were processed as per the flow chart shown in figure 1.
Figure. 1. Flow Chart of the Proposed Method
Initially pre-processing is performed to suppress unwanted distortions or enhance image features important for further processing. In this study, we have applied median filter to remove the noise. The Median Filter is a non-linear digital filtering technique, often used to remove noise from a picture or signal. Median filtering is widely utilized in digital image processing as it preserves edges while removing noise. The Median filter is a sliding-window spatial filter. It replaces the value of centre pixel with the median of the intensity values in the neighborhood of that pixel to reduce “salt and pepper” noise [14] . As thermal images are already corrected for thermal noise by inbuilt software in thermal cameras elimination of noise by filters is not required.
Next the eye region is segmented from the overall thermal image of the face. This region of interest is obtained by manual cropping of a rectangular region, including both eyes, while automatic segmentation[15] is being worked on. In order to study the distribution of temperature in the eye, contours were extracted. Contour is a curve joining all the continuous points having same color or intensity. The contours are a useful tool for shape analysis, object detection and recognition. Here line contours are extracted to detect the points of same temperature and hence study temperature distribution patterns in normal and abnormal eyes.
III. RESULTS AND DISCUSSION
Ocular thermograms were obtained from all volunteers in accordance to the thermal image acquisition protocol. As the camera is placed at a distance of 1 meter from the subject, the heat profile of the entire face is captured on the thermogram. A rectangular region containing both eyes is segmented. Contours are obtained from this region of interest. The thermal image of face, the segmented ROI and Contours obtained are shown for all cases in figures a,b, c respectively. As a preliminary study, the shape, size and symmetry of these contours were manually analyzed in all normal volunteers and deviations were looked for in the abnormal cases.
Firstly the ocular thermograms obtained from 19 normal subjects were processed and analysed. One sample case (N008) from the normal volunteers has been presented in CASE 1 with necessary figures for illustration purpose.
(a) (b) ( c )
Figure 2. Normal Study
Fig. 2(a) shows the thermal image of complete face acquired from the normal volunteer.
Fig. 2(b) shows the eye region segmented from the thermal image of face. The contour profile extracted from the segmented eye is shown in Fig. 2(c). It is observed that pupil is defined as circle in the centre. Also the thermal contours are distributed in as concentric circles around the pupil. This pattern was observed for all the 19 normal persons, thereby establishing a distinct thermal distribution pattern for normal eyes.
Secondly, ocular thermograms acquired from five volunteers with different abnormalities in vision were analyzed. These are presented in case studies 2 – 6. In Case 2, ocular thermogram was acquired from a 20 year male with eye power of 3.57. The subject regularly uses spectacles for vision correction. The thermal image was acquired with no spectacles in order to study the pattern deviations that may be caused by the vision abnormality. Figure 3 shows the thermal image of face, ROI and contours. It is observed that the pupil contour is not defined perfectly as in the normal cases. The concentric contours are also not clearly defined.
(a) (b) ( c )
Figure3. Abnormal Study - Eyes with vision defect
(a) (b) ( c )
Figure 4. Abnormal Study - Eyes with Cataract
Case 3 was a 67 year old female diagnosed with cataract problem. The thermal image of face, ROI and contours are shown in figure 4. In this case, the thermal patterns are not clearly visible. Pupil and thermal contours were not clear. In Case 4, thermal contours were extracted from a twenty year old insomniac male as displayed in figure 5. It was observed that there were no significant deviations from normal contour patterns. But a significant elevation in eye temperature was noted.
(a) (b) ( c ) Figure 5. Abnormal Study - Insomniac Eyes
Case 5 was a 53 year old male who had reported cataract and undergone cataract surgery before 8 years. In addition he was presently diagnosed with diabetes mellitus. The contours obtained were significantly different from the normal pattern as seen in figure 6. The central circle of pupil and iris were not defined and closed concentric patterns were not present.
(a) (b) ( c ) Figure 6. Abnormal Study - diabetic Eyes
Case 6 was a 48 year old male with Presbyopia. It is an aging related condition that results in gradual loss of eyes' ability to focus on nearby objects. Figure 7 shows the thermal images and contours obtained from this abnormal volunteer. In this case, the pupil region is observed to be larger in size compared to normal while the concentric thermal contours remained unchanged.
(a) (b) ( c ) Figure 7. Abnormal Study - Presbyopia
A case study was conducted to observe the effect of prolonged usage of mobile phones in students. A 20 year old male student was asked to watch videos on mobile phone for 10 minutes, keeping full brightness setting for the display. The thermal images of his eyes were acquired before and after the long exposure to mobile phone display. The images are shown in figure 8.
(a) Original (b) Before ( c ) After
Figure 8. Effects of Prolonged use of Mobile Phones
Initially, the average eye temperature was measured to be 34oC. After the mobile challenge an increase in 3oC was observed.
IV. CONCLUSION
In this paper, image processing techniques have been used to study the thermal patterns in normal and abnormal eyes. In the normal eye, the pupil is defined as a closed circle and thermal contours were found to be distributed in concentric closed curves around the pupil. In the abnormal eyes, deviations from this pattern has been detected by visual observation. From this preliminary case study with 25 subjects, it is inferred that the thermal patterns of normal and abnormal eyes are significantly distinct. Also the study confirms that prolonged use of illuminated displays as with mobile phones results in an increase in eye temperature. Further study involving a larger group of controlled subjects may help to validate these preliminary findings. Also image features which can clearly discriminate between normal and abnormal curved patterns in the ocular thermal images may be extracted to develop an automatic screening technique for early detection of eye abnormalities.
V. ACKNOWLEDGEMENT
We would like to thank Dr. Sandeep Jaipurkar,Consultant Radiologist, Vijaya Health Centre, Vadapalani, Chennai, for providing the thermal ocular image acquisition facility and valuable clinical validation for this study.
VI. REFERENCES
[1] Hassan Hashemi, MD, Mohammad Saatchi, MS, AbbasaliYekta, MS, Babak Ali, MS, HadiOstadimoghaddam, PhD, PayamNabovati, PhD, MohamadrezaAghamirsalim, MD, and Mehdi Khabazkhoob, PhD , „ High Prevalence of Asthenopia among a Population of University Students‟ Journal of Ophthalmic and Vision Research. 2019 ; 14(4): 474–482.
[2]Takashi Kojima, Motoko Kawashima, Shigeru Nakamura, KazuoTsubota, „Advances in the diagnosis and treatment of dry eye‟ , Progress in Retinal and Eye Research, 2020; Volume 78.
[3]Takashi Itokawa, Yukinobu Okajima, Takashi Suzuki, Tatsuhiko Kobayashi, YutoTei, Koji Kakisu, and Yuichi Hori , „Association among Blink Rate, Changes in Ocular Surface Temperature, Tear Film Stability, and Functional Visual Acuity in Patients after Cataract Surgery‟, Dry Eye Disease and Refractive Corrections, 2019 |Article ID 8189097 | https://doi.org/10.1155/2019/8189097
[4]Jen-HongTan, E.Y.K.Ng, U.Rajendra Acharya, C.Chee, „Infrared thermography on ocular surface temperature: A review‟, Infrared Physics & Technology
Volume 52, Issue 4, July 2009, Pages 97-108
[5] PadmapriyaNammalwar, VenkateswaranNarasimhan ,Toshita Kannan and Sindhu MadhuriMorapakala, “Noninvasive Glaucoma Screening Using Ocular Thermal Image Classification” ,CIT. Journal of Computing and Information Technology, Vol. 25, No. 3, September 2017, 227–236.
[6] D. Selvathi and K. Suganya, "Support Vector Machine Based Method for Automatic Detection of Diabetic Eye Disease using Thermal Images," 2019 1st International Conference on Innovations in Information and Communication Technology (ICIICT), Chennai, India, 2019, pp.
1-6doi: 10.1109/ICIICT1.2019.8741450.
[7] M. NaeemHussien, Mohd-Haris Lye, Mohammad Faizal Ahmad Fauzi, Tan ChingSeong and SarinaMansor, “Comparative Analysis of Eyes Detection on FaceThermal Images”, Proc. of the 2017 IEEE International Conference on Signal and Image Processing Applications (IEEE ICSIPA 2017), Malaysia ,September 12-14, 2017.
[8] Tai Yuan Su, Chen KerhHwa, Po Hsuan Liu, Ming Hong Wu, David O. Chang, Po Fang Su, Shu Wen Chang, and Huihua Kenny Chianga, “Noncontact detection of dry eye using a custom designed infrared thermal image system”, Journal of Biomedical Optics 16(4), 046009 (April 2011)
.[9]B.B. Lahiri, S. Bagavathiappan, T. Jayakumar, and John Philip⁎Medical applications of infrared thermography: A review, Infrared Phys Technol. 2012 Jul; 55(4): 221–235.
[10] E F J Ring and K Ammer, Infrared thermal imaging in medicine,Physiological Measurement, 2012, Volume 33, Number 3.R 33-46.
[11] SheejaVFrancis, M.Sasikala, G.BhavaniBharathi, Sandeep D.Jaipurkar, „Breast cancer detection in rotational thermography images using texture features‟,Infrared Physics & Technology, Volume 67, November 2014, Pages 490-496.
[12] U. Snekhalatha, M.Anburajan, T. Teena, B. Venkatraman, M. Menaka and B. Raj, "Thermal image analysis and segmentation of hand in evaluation of rheumatoid arthritis," 2012 International Conference on Computer Communication and Informatics, Coimbatore, India, 2012, pp. 1-6, doi: 10.1109/ICCCI.2012.6158784.
[13] https://infraredcameras.com/thermal-infrared-products/ici-p-series-ir-camera-8640/
[14] Gonzalez, R. C., & Woods, R. E. (2002). Digital Image Processing (6th edition).
[15] Rafika HARRABICEREP Research Unit, Ezzedine BEN BRAIEK CEREP Research Unit,“Color Image Segmentation Using a Modified Fuzzy C-Means Technique and different color spaces: Application in the Breast Cancer Cells Images”, 1st International Conference on Advanced Technologies for Signal and Image Processing - ATSIP'2014 March 17-19, 2014, Sousse, Tunisia.
