Information Technology Reference
In-Depth Information
the coating color had turned to purple. He published a topic about his study,
including in it 130 tongue inspection images, along with gastroendoscopy images
and patient diagnoses [20].
Since the early 1980s, medical professionals in China have systematically
studied the relationship between various cancers and tongue signatures. Their
results have been published in national medical journals. For instance, China
TCM Society, China Cancer Society and TCM Diagnosis Association conducted
a national project that included cases of 12,448 cancerous patients, 1,628 non-
cancerous patients and 5,578 normal patients. The results statistically showed
that there are significant changes of color, coating, shape and dorsum shape of
the tongues of cancerous patients versus those tongues of non-cancerous patients
or normal subjects [23]. Fujin TCM Hospital in China conducted a survey of 168
stomach cancer patients and 200 healthy subjects in 1983. The results showed
that the percentage of abnormal tongues were 4 to 10 times higher in cancerous
patients than those of healthy subjects; for example, for cancer patients: 70.8%
purple/bluish color, 76.5% abnormal proper coating, 20.8% with cracks, and
83.9% dorsum deformation, and for healthy subjects: only 12.5% purple/bluish
color, 26.2% abnormal proper coating, 1.5% with cracks, and 10.5% with dorsum
deformation [25]. PLA 211 Hospital in China conducted a series of surveys on
liver cancer patients versus non-cancer patients who have liver diseases. Results
showed that liver cancer patients' abnormal tongue percentage is about 2 to
3 times higher than those of non-cancerous patients [21]. The data shows the
method has certain selectivity for cancer diagnosis. Another survey showed that
the level of abnormal signatures on the tongue increased as the tumor size in-
creased from less than 5mm, larger than 5mm, to wide spread. The data shows
promise for cancer stage estimation and possible early diagnosis [63].
Visual inspection of the tongue offers many advantages: it is a non-invasive
diagnosis method, is simple and inexpensive. However, the current practice in
TCM is mainly experience based or subjective. The quality of the visual inspec-
tion varies between individuals. Although there are a few experts successfully
diagnosing cancers based on inspection of the tongue, their skills are not easily
transferable to other medical professionals. Their expertise is limited to qualita-
tive descriptions, not to quantitative or mathematical formulations. To circum-
vent this problem, studies have investigated and reported options such as fuzzy
logic [30] and image analysis [31]. Here we discuss a computerized vision system
for tongue inspection.
In this study, we investigate a novel imaging system for visual inspection of
the tongue [19]. The objectives are to use a digital camera to make an image
of a patient's tongue, then use software to extract the features from the digital
image created, and finally make a diagnosis based on quantitative models. The
goal is not to replace the conventional diagnostic methods but to give an early
alert signal that can lead to further diagnosis by other methods, such as MRI,
CT, X-ray, etc. This novel approach has various significant advantages. First, it
makes the inspection objective and repeatable so that it prevents human bias
and errors. Second, it can be implemented on an inexpensive personal computer
Search WWH ::




Custom Search