Abstract: 3D scanner captures the appearance and geometry of a real object, and forms a virtual one which can be displayed in 3D on computers. However, the high cost of 3D scanning devices inhibits the popularity of scanned 3D objects on the web and related applications. Our system targets for developing a consumer-level 3D scanning device suitable for naïve web users with the use of a novel natural depth sensing device – Kinect. As the Kinect device is a low-cost, robust and fast depth camera with color, these essential features enable us to acquire RGBD images from different views of the target object. To facilitate the registration of point clouds from various depth images, we first combine marker-based tracking system to estimate the current view angle of Kinect during capture. Then, we employ the Iterative Closest Point (ICP) algorithm with the efficiency feature detection to reconstruct a single point cloud representing the 3D object surface. To reduce size of point cloud, a down-sampling is applied to remove redundant and densely sampled points in the model. The whole process of capturing, storing, uploading and displaying 3D data on web application is done with minimal user involvements. Users can freely control the viewing angle of their favorite product with the 3D scanned point cloud object displayed in the browser; this greatly enhances user's shopping experience. We believe that the introduction of attractive and fashionable 3D product on the web will gain significant attention from customers and evolve habits and traditions in e-commerce, especially for C2C business, e.g. auction sites like ebay and taobao, in which sellers can promote their goods for sale with scanned 3D models