Termal görüntü analizi ile tümörlerin erken tanılanması
Yükleniyor...
Tarih
2022
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Ege Üniversitesi, Fen Bilimleri Enstitüsü
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Günümüzde, tümörlerin sezimlenmesinde kullanılan birçok cihaz bulunmaktadır. Ancak bu cihazlar çok pahalı olup, küçük çaplı tümörlerin sezimlenmesinde yetersiz kalırken, hastaya sıkıntı verici bir uygulama olup, aynı zamanda iyonize radyasyonun zararlı etkisini beraberinde getirebilmektedir. Diğer yandan, termal görüntüleme; çok düşük maliyetli ve yan etkisiz bir yöntem olup, erken dönem kanserin tanısında geleneksel yöntemlere göre yüksek doğruluğu ile önemli bir aday yöntem olarak görülmektedir. Bu çalışma meme tümörlerinin tanılanması için üç adımdan oluşmaktadır. İlk adım; Comsol Multiphysics benzetim programı ve programın Bio-ısı Transfer Modülü kullanılarak sağlıklı meme dokusu ve tümör dokusu taklit edilerek asimetri durumları oluşturulmasıdır. İkinci adım; içine tümör taklidi yapan dirençler gömülerek oluşturulan meme fantom ortamı tasarımı ve meme fantomundan çeşitli asimetri senaryoları oluşturularak termal kamera ile görüntü almaktır. Üçüncü adım; veritabanından alınan, fantom yapay organından elde edilen ve benzetim programı ile üretilen görüntülerin bölütlenmesi ve nicemsel olarak yapılan piksel parlaklığı ve damarlaşma (veritabanı) tabanlı asimetri çalışmalarıdır. Meme bölgesi sağ ve sol meme olarak otomatik olarak bölütlenmiştir. Bölütlenen görüntülerden piksel parlaklığı tabanlı öznitelikler çıkartılmıştır. Bölütlenen termal görüntüler (veritabanı) kullanılarak damar bölütlemesi yapılmıştır. Daha sonra damarlaşma tabanlı öznitelikler çıkartılmıştır. Bu öznitelikerin farklarının standart sapması ve ortalaması hesaplanarak piksel parlaklığı ve damarlaşma (veritabanı) tabanlı asimetri analizleri yapılmıştır. Fantom ve Comsol Multiphysics programı ile elde edilen görüntülerin analizi sonucunda sağlıklı doku ve tümörlü doku arasında sıcaklık (piksel parlaklığı) olarak asimetrilerin oluştuğu tespit edilmiştir. Bu sebeple termal görüntülemenin tümörlerin erken tanılamada tamamlayıcı bir araç olduğunu söyleyebiliriz.
Today, there are many devices used in the detection of tumors. However, these devices are very expensive and insufficient for detecting small-scale tumors, and they have some limitations due to patient discomfort associated with compression of the breast tissue and exposure to ionizing radiation. On the other hand, thermal imaging is a very low-cost and side-effect-free method, and it is seen as an important candidate method in the diagnosis of early-stage cancer with its high accuracy compared to conventional methods. In this study there are three steps for the diagnosis of breast tumors. The first step is to create asymmetry conditions by simulating healthy breast tissue and tumor tissue using the Comsol Multiphysics simulation program and it’s Bio-heat Transfer Module. The second step is to design the breast phantoms, which are created by embedding resistors that mimic a tumor, and take images with a thermal camera by creating various asymmetry scenarios from the breast phantom. The third step is segmentation, and quantitative intensity value and vascularization (database)-based asymmetry analyses of the thermal images. These are updated from the database, the phantom artificial organ and the simulation program. Firstly, the breast region was automatically segmented as right half breast and left half breast. Then, intensity value based features were extracted from the segmented images. Next, vessel segmentation was performed using segmented thermal images (database). After that, the vascular quantification based features were extracted. Finally, intensity value and vascular quantification (database) based asymmetry analyses were performed by calculating the standard deviation and average of these features differences. As a result of the analysis of the images obtained from the Phantom and Comsol Multiphysics program, it was determined that asymmetries in terms of temperature (pixel intensity) occurred between the healthy tissue and the tumor tissue. Hence, we can say that thermal imaging is a complementary tool in the early diagnosis of breast tumors.
Today, there are many devices used in the detection of tumors. However, these devices are very expensive and insufficient for detecting small-scale tumors, and they have some limitations due to patient discomfort associated with compression of the breast tissue and exposure to ionizing radiation. On the other hand, thermal imaging is a very low-cost and side-effect-free method, and it is seen as an important candidate method in the diagnosis of early-stage cancer with its high accuracy compared to conventional methods. In this study there are three steps for the diagnosis of breast tumors. The first step is to create asymmetry conditions by simulating healthy breast tissue and tumor tissue using the Comsol Multiphysics simulation program and it’s Bio-heat Transfer Module. The second step is to design the breast phantoms, which are created by embedding resistors that mimic a tumor, and take images with a thermal camera by creating various asymmetry scenarios from the breast phantom. The third step is segmentation, and quantitative intensity value and vascularization (database)-based asymmetry analyses of the thermal images. These are updated from the database, the phantom artificial organ and the simulation program. Firstly, the breast region was automatically segmented as right half breast and left half breast. Then, intensity value based features were extracted from the segmented images. Next, vessel segmentation was performed using segmented thermal images (database). After that, the vascular quantification based features were extracted. Finally, intensity value and vascular quantification (database) based asymmetry analyses were performed by calculating the standard deviation and average of these features differences. As a result of the analysis of the images obtained from the Phantom and Comsol Multiphysics program, it was determined that asymmetries in terms of temperature (pixel intensity) occurred between the healthy tissue and the tumor tissue. Hence, we can say that thermal imaging is a complementary tool in the early diagnosis of breast tumors.
Açıklama
Anahtar Kelimeler
Termogram, Meme Tümörü Tespiti, Asimetri Analizi, Damarlaşma Analizi, Meme Bölgesinin Bölütlenmesi, COMSOL Multiphysics, Thermogram, Breast Cancer Detection, Asymmetry Analysis, Vascular Quantification Analysis, Breast Segmentation, COMSOL Multiphysics
