Abstract:We present a novel fully probabilistic method to interpret a single face image with the 3D Morphable Model. The new method is based on Bayesian inference and makes use of unreliable image-based information. Rather than searching a single optimal solution, we infer the posterior distribution of the model parameters given the target image. The method is a stochastic sampling algorithm with a propose-and-verify architecture based on the Metropolis–Hastings algorithm. The stochastic method can robustly integrate unreliable information and therefore does not rely on feed-forward initialization. The integrative concept is based on two ideas, a separation of proposal moves and their verification with the model (Data-Driven Markov Chain Monte Carlo), and filtering with the Metropolis acceptance rule. It does not need gradients and is less prone to local optima than standard fitters. We also introduce a new collective likelihood which models the average difference between the model and the target image rather than individual pixel differences. The average value shows a natural tendency towards a normal distribution, even when the individual pixel-wise difference is not Gaussian. We employ the new fitting method to calculate posterior models of 3D face reconstructions from single real-world images. A direct application of the algorithm with the 3D Morphable Model leads us to a fully automatic face recognition system with competitive performance on the Multi-PIE database without any database adaptation.

A Hybrid Face Recognition Method Using Markov Random Fields

Towards Mask-robust Face Recognition.

Face alignment based on high order markov random field

Face Alignment Based on 3D Face Shape Model and Markov Random Field.

Combination of Local Multiple Patterns and Exponential Discriminant Analysis for facial recognition

A Hybrid Face Recognition Algorithm Based on WT, NMFs and SVM

Face Recognition Using Survival Exponential Entropy: Based on Markov Random Field Modeling

Face Recognition with Contiguous Occlusion Using Markov Random Fields

Multi-Pose Face Recognition Using Hybrid Face Features Descriptor

Markov Chain Monte Carlo for Automated Face Image Analysis

Learning Dynamic Hybrid Markov Random Field for Image Labeling

Face Super-Resolution Using 8-Connected Markov Random Fields with Embedded Prior

Patch-based face recognition using a hierarchical multi-label matcher

Combining Local Descriptions With Geometric Constraints For 3-D Object Recognition In Multiple Statuses

Parts-Based holistic face recognition with RBF neural networks

HiMFR: A Hybrid Masked Face Recognition Through Face Inpainting

A novel face recognition system using hybrid neural and dual eigenspaces methods

A hybrid approach for face recognition using a convolutional neural network combined with feature extraction techniques

Multiscale collaborative representation for face recognition via class-information fusion

A Face Recognition Method Based on Fuzzy Data Fusion

Hybrid Approach to Face Recognition System using Principle component and Independent component with score based fusion process