HyperAI超神经

首页资讯最新论文教程数据集百科 SOTA LLM 模型天梯 GPU 天梯顶会

中文

HyperAI超神经

3D Face Reconstruction On Now Benchmark 1

评估指标

Mean Reconstruction Error (mm)

Median Reconstruction Error

Stdev Reconstruction Error (mm)

评测结果

各个模型在此基准测试上的表现结果

模型名称	Mean Reconstruction Error (mm)	Median Reconstruction Error	Stdev Reconstruction Error (mm)	Paper Title	Repository
RingNet	1.53	1.21	1.31	Learning to Regress 3D Face Shape and Expression from an Image without 3D Supervision
DECA	1.38	1.09	1.18	Learning an Animatable Detailed 3D Face Model from In-The-Wild Images
DenseLandmarks (Single-view)	1.28	1.02	1.08	3D face reconstruction with dense landmarks	-
PIXIE	1.49	1.18	1.25	Collaborative Regression of Expressive Bodies using Moderation
PRNet	1.98	1.50	1.88	Joint 3D Face Reconstruction and Dense Alignment with Position Map Regression Network
FOCUS	1.30	1.04	1.10	Robust Model-based Face Reconstruction through Weakly-Supervised Outlier Segmentation
FLAME template	1.53	1.21	1.31	Learning a model of facial shape and expression from 4D scans
Dib et al. 2021	1.57	1.26	1.31	Towards High Fidelity Monocular Face Reconstruction with Rich Reflectance using Self-supervised Learning and Ray Tracing	-
3DDFA_V2	1.57	1.23	1.39	Towards Fast, Accurate and Stable 3D Dense Face Alignment
UMDFA	1.89	1.52	1.57	“Look Ma, no landmarks!” – Unsupervised, Model-based Dense Face Alignment
SynergyNet	1.59	1.27	1.31	Synergy between 3DMM and 3D Landmarks for Accurate 3D Facial Geometry
3DMM-CNN	2.33	1.84	2.05	Regressing Robust and Discriminative 3D Morphable Models with a very Deep Neural Network
Deng et al. 2019	1.54	1.23	1.29	Accurate 3D Face Reconstruction with Weakly-Supervised Learning: From Single Image to Image Set
MICA	1.11	0.90	0.92	Towards Metrical Reconstruction of Human Faces
MGCNet	1.87	1.31	2.63	Self-Supervised Monocular 3D Face Reconstruction by Occlusion-Aware Multi-view Geometry Consistency
DenseLandmarks (Multi-view)	1.01	0.81	0.84	3D face reconstruction with dense landmarks	-
Deep3DFaceRecon PyTorch	1.41	1.11	1.21	Accurate 3D Face Reconstruction with Weakly-Supervised Learning: From Single Image to Image Set

0 of 17 row(s) selected.