Ứng dụng phương pháp xử lý ảnh trong hỗ trợ phát hiện da beo ở cổ tử cung

Người thực hiện: Nguyễn Ngọc Quỳnh

Tóm tắt kết quả:

In the present study, based on the distribution of blood and glycogen in normal cervical tissue and leopard-skin cervical tissue, we developed an automatic method of leopard skin detection using a digital colposcopy image. The results of image segmentation were compared and evaluated by the traditional method – cervical image after applying Lugol’s solution. Pictures before and after applying Lugol’s solution were taken at different times. Therefore, the evaluation results are mainly based on the shape of the inflammation point distribution on the segmental image corresponding to the shape of the distribution of inflammation points on the image after applying Lugol. The research opens a new approach to support doctors in quickly diagnosing diseases, making conclusions about the condition, and appropriate treatment options. At the same time, it solves the limitations of traditional methods. However, further studies need to be done to assess method sensitivity and specificity.