In this project, using deep learning techniques, we tried to detect the need for Apicoectomy surgery through dental Orthopantomography images. We have used U-NET. Also for feature extraction from DenseNet169 Used.
This project was done by Alpha team (Mojtaba Zarreh, Mohammad Salar Cheraghi).
When the soft inner pulp of your tooth becomes infected, there is only one way to treat the problem and save the tooth There is your natural and that is to remove the pulp from the tooth.
Dentists can usually fix this problem and restore the tooth with the help of non-surgical (non-invasive) methods or root canal treatment, but sometimes there is no other option than dental surgery.
When the treatment is done, the pulp is pulled out through the outer crown of your tooth.
When direct access to the pulp and root of the tooth is needed, an incision is made on the gum.
This method helps the dentists to clean the pulp and infection through the root of the tooth and to remove the infected tissues of the bone under the gums.
An apicoectomy (also referred to as apicectomy, root end resection or periapical surgery) is a minor dental procedure that is performed to save a tooth that would otherwise need to be removed. An apicoectomy involves removing the end of the root of a tooth (known as the apex) and is performed on both children and adults.
We are trying to diagnose the need for this surgery in OPG modality.
The dataset contains a limited number of whole-mouth OPG images labeled by experienced dentists.
It also contains images of masks for problematic teeth.
The image format is DICOM.
https://www.kaggle.com/datasets/iaaaevent/iaaa-v2
Abnormal images with masks applied
As mentioned above, we have used the U-NET architecture.
And DenseNet169 has been used for feature extraction.
The final architecture of the model :