| Proj No. | A3104-251 |
| Title | BRAIN TUMOR DETECTION: TRADITIONAL VS. AI-BASED APPROACHES |
| Summary | Medical experts use the MRI/CT scans for various types of cancer detection; however, a challenging problem in MRI/CT data is the presence of noise, which can lead to misdiagnosis. Over the years, several algorithms have been developed, which provide reasonable results under certain conditions and fail in other situations. Thus, the purpose of this project is to compare traditional and AI/ML based algorithms for brain tumour detection. Traditional Image Processing Approach will include the following steps. ✔ Preprocessing: Convert to grayscale, apply Gaussian filtering to remove noise. ✔ Edge Detection: Use Sobel or Canny edge detection to highlight tumours regions. ✔ Segmentation: Apply Otsu’s Thresholding & Watershed Algorithm to segment the tumor. ✔ Feature Extraction & Classification: Use SVM (Support Vector Machine) or KNN (K-Nearest Neighbors) for tumor classification. AI/ML Approach will include the following steps: ✔ Use a Convolutional Neural Network (CNN) to automatically extract tumor features. ✔ Train on labeled datasets (e.g., BraTS 2020 Dataset). ✔ Use Transfer Learning (VGG16, ResNet, InceptionNet) for improved accuracy. During the final stage of this project, the result of these algorithms will be analysed and compared in order to choose the best algorithm for medical applications. |
| Supervisor | A/P Mohammed Yakoob Siyal (Loc:S2 > S2 B2A > S2 B2A 28, Ext: +65 67904464) |
| Co-Supervisor | - |
| RI Co-Supervisor | - |
| Lab | Computer Engineering I (Loc: S2-B4c-15) |
| Single/Group: | Single |
| Area: | Digital Media Processing and Computer Engineering |
| ISP/RI/SMP/SCP?: |