Construction and Reconstruction of 3D Facial and Wireframe Model Using Syntactic Pattern Recognition
Shilpa Rani, Deepika Ghai, Sandeep Kumar · 2021
Day by day, the entire world depends on human-computer interactive machines because demands increase in real-time devices being a security concern. The human-computer interactive machine will make human life easy, fast, and more stable. Face recognition is one of the popular applications of image analysis because it has a wide range of law and commercial applications and improved technology after 30 years of continuous research. Human face recognition using machines continuously attracting the researchers and gaining the machine learning–based recognition system gained a certain level of maturity with some limitations. Recognition of face in different illumination and pose variation is still a mostly unsolved problem. 3D face model– based methods are providing the solution to these problems. 3D face models can be designed using image-based methods or laser scan-based methods. Image-based face modeling methods are the more significant reconstruction of the model done through one or more 2D image slices. Our goal is to create an algorithm that can be used for the reconstruction of a 3D model. In this paper, we proposed a novel approach to reconstructing 3D face models using 2D images. First, the facial features are extracted using the syntactic pattern recognition technique. In the second step, dense features are extracted using SIFT technique. In the third step, the shape and shading technique is applied to extract the image's enhanced details. In the fourth step, the Basel face model is applied to create the 3D face image. Our proposed method is efficient because it requires a single image to reconstruct a 3D face model and it is also reconstructs the wireframe model of the image if the input image is a wireframe model. The proposed methodology's efficiency is verified on publicly available benchmark databases, i.e., Face Warehouse database, USF dataset, and our dataset. The proposed methodology's effectiveness is represented in the result section by comparing the proposed methodology's RMS value with existing methods. The proposed method can design the facial mask for facial burn victims from pre-burn photos.