Atexture Classification Using Random Forest And Decision Tree
Mohammed M. Razooq · VAWKUM Transactions on Computer Sciences · 2015
Texture analysis is considered fundamental and important in the fields of pattern recognition, computer vision and image processing. Analysis of the textures involves texture features extraction and selection, and plays an essential role in the classification and segmentation of textural features. In this study have compared two texture classification methods based on the Random Forest (RF) and Decision Tree (DT) classifiers by using a combination method between various extraction features, such as bi-orthogonal wavelet transform, gray level histogram and edge detection. Experiments were conducted on two different databases. The first texture database captured digital images for testing multi-class machining processes, and the second database was collected from the Brodatz album. The results have revealed that RF and DT have yielded higher classification precision.