A Comparative Study of Clustering Approaches on Segmentation for Construction Remodeling

Discovery of a new clustering technique for remodeling building structures is more advantageous for the consumers and construction Engineers. A Comparative study is required to different clustering approaches is required for the application of Remodeling. Analysing various clustering techniques will invoke a new idea for this research. The various clustering techniques are approached to the segmented image of dented building structures and to rejoin them to remodel. In this work, the parallel mean shift clustering technique and fuzzy C-means clustering are applied to similar datasets and the outcome are compared. The result is compared with spectral clustering method. This work guides the civil engineers to remodel the existing damaged building segments. The best segments are selected to rejoin by any suitable clustering technique. Then the complexity of results is verified with the similar metrics probably to find the better clustering method. In Parallel Mean shift clustering, the unlabelled training data is used and hence it forms a unsupervised learning algorithm. Whereas the fuzzy C means clustering takes m x n two dimensional matrix formats which maps the area and volume of the building segments.