Hybrid Learning Method for Image Segmentation

Image segmentation by graph partitioning is popular in the field of artificial intelligence and computer vision, so it is the subject of several researches due to the good performance in a wide range of applications. Many image segmentation techniques employ classical machine learning processes and extract features according to machine-learning methods whereas classifying features via highly specialized training programs. It is becoming, an important branch of artificial intelligence and computer science. This paper put forward a new method based on traditional machine learning, which combines Random forest with the problem of Image Segmentation by Graph Partitioning. This paper introduced a novel clustering algorithm based on a Graph cut generated with a random forest. We test our method on the dataset BRATS, some lungs Images, and standard test image Lenna.