Exploring the Role of Extracted Features in Deep Learning-based 3D Face Reconstruction from Single 2D Images

In 3D face reconstruction, features extracted from a single image are crucial as they provide extra information beyond the image’s size and quality. These features, including facial landmarks and face parsing, texture and UV maps, depth maps, shading information, and albedo maps, can be used to compensate for the lack of prior knowledge provided by a single 2D image and to overcome the dimensional differences between 2D and 3D. By using these features, 3D face reconstruction neural networks can create more detailed and accurate 3D models of faces, even when the input image is of low quality or has extreme poses or occlusions. The paper investigates the definition, importance, extraction, and various applications of each feature, including their potential use in 3D face reconstruction and facial recognition. The paper will provide an analysis of the current state of the art in this field and discuss the potential for future research and development.