皮膚淋巴癌為一種罕見的肌膚疾病,由於其發生的機率並不高,因此若非專業皮膚科醫生進行診療,容易被誤診為常見的乾癬或是異位性皮膚炎。傳統活體切片檢測屬侵入式過程,除破壞組織結構外,其病理診斷所需費用較高且費時。我們使用了多頻譜影像技術,得到肌膚疾病患者影像中每個像素的頻譜,並利用主成份分析與主成份分數圖以分辨出不同的疾病。並利用4波段LED的色差計算,找出最能夠有效增益色差之LED色度位置,將其與白點連線,可得一個特定波長。利用此波長,我們設計了一個增益頻譜,透過此增益頻譜,我們可以有效增益皮膚淋巴癌患者影像中,正常與病變區域之色差(增加10 %以上),能有效地提升醫生診斷效率,幫助病患進行早期治療。
Cutaneous T-cell Lymphoma is a rare skin disease. Due to the low probability of its occurrence, non-professional dermatologist clinics could easily be mistaken for psoriasis or atopic dermatitis. Traditional biopsy detection is an invasive process, it not only undermines the organizational structure, and the cost of pathological diagnosis is high and time-consuming. We used multi-spectral imaging technique to get the spectrum of each pixel in the image of these skin diseases, and use of principal component analysis and principal component scores to distinguish between different diseases. Later, we use 4-band LED color difference calculation to find out the best able to gain color LED chromaticity location, its connection with the white spots can be a particular wavelength. By using this wavelength, we designed an enhancing spectrum, and then we can enhance the color difference between normal and lesion area of the CTCL patients more than 10%. It can make medical diagnosis more efficiency and help patients for early treatment.