Neural Face Editing with Intrinsic Image Disentangling SUPPLEMENTARY MATERIAL
نویسندگان
چکیده
1. Implementation: more details In this section, we provide more details regarding the implementation of the rendering layers fshading and fimage-formation as described in the paper. 1.1. Shading Layer The shading layer is rendered with a spherical harmonics illumination representation [6, 2, 7, 1]. The forward process is described by equations (3),(4), and (5) in the main paper. We now provide the backward process, i.e., the partial derivatives ∂S i e ∂nx , ∂S i e ∂ny , ∂S i e ∂nz and ∂S i e ∂Lj as follows: ∂S e ∂nx = 2(c1L9nx + c1L5ny + c1L8nz + c2L4) (1)
منابع مشابه
Supplementary Material: Intrinsic Face Image Decomposition with Human Face Priors
In Fig. 2–Fig. 14, we present additional comparisons to the current state-of-art singleimage methods for face modeling [2] and intrinsic image decomposition [1]. We show results with white illumination in Fig. 2–Fig. 7 and non-white illumination in Fig. 8– Fig. 14. Note the differences in decomposed albedo maps between the Caucasian woman in Fig. 13 and the black woman in Fig. 14 under the same...
متن کاملLabel Denoising Adversarial Network (LDAN) for Inverse Lighting of Face Images
Lighting estimation from face images is an important task and has applications in many areas such as image editing, intrinsic image decomposition, and image forgery detection. We propose to train a deep Convolutional Neural Network (CNN) to regress lighting parameters from a single face image. Lacking massive ground truth lighting labels for face images in the wild, we use an existing method to...
متن کاملIntrinsic Face Image Decomposition with Human Face Priors
We present a method for decomposing a single face photograph into its intrinsic image components. Intrinsic image decomposition has commonly been used to facilitate image editing operations such as relighting and re-texturing. Although current single-image intrinsic image methods are able to obtain an approximate decomposition, image operations involving the human face require greater accuracy ...
متن کاملMulti-View Face Detection in Open Environments using Gabor Features and Neural Networks
Multi-view face detection in open environments is a challenging task, due to the wide variations in illumination, face appearances and occlusion. In this paper, a robust method for multi-view face detection in open environments, using a combination of Gabor features and neural networks, is presented. Firstly, the effect of changing the Gabor filter parameters (orientation, frequency, standard d...
متن کاملAn Adversarial Neuro-Tensorial Approach For Learning Disentangled Representations
Several factors contribute to the appearance of an object in a visual scene, including pose, illumination, and deformation, to mention a few. Each factor accounts for a source of variability in the data, while the multiplicative interactions of these factors emulate the entangled variability, giving rise to the rich structure of visual object appearance. Disentangling such unobserved factors fr...
متن کامل