N Heads Are Better Than One: Exploring Theoretical Performance Bounds of 3D Face Reconstruction Methods

William Rowan; Patrik Huber; N. E. Pears; Andrew Keeling

N Heads Are Better Than One: Exploring Theoretical Performance Bounds of 3D Face Reconstruction Methods

William Rowan, Patrik Huber, N. E. Pears, Andrew Keeling

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

We introduce ``N Heads Are Better Than One'', a novel approach for evaluating combinations of existing 3D face reconstruction methods. By calculating lower theoretical error bounds for method combinations on the NoW benchmark, we establish a robust set of new baselines for the task of 3D face reconstruction. Our work also provides a framework for assessing the potential of these aggregate `pseudo-foundation models,' which leverage strengths from multiple existing approaches. In doing so, we improve understanding of the performance of current methods and set targets for future foundation models to beat.

Original language	English
Title of host publication	European Conference on Computer Vision Workshop (ECCVw) 2024
Subtitle of host publication	Foundation Models for 3D Humans
Publisher	Springer Science + Business Media
Number of pages	9
Publication status	Published - 4 Oct 2024

Bibliographical note

This is an author-produced version of the published paper. Uploaded in accordance with the University’s Research Publications and Open Access policy.

Keywords

3D Face Reconstruction; Foundation Models; Performance Evaluation

Access to Document

Foundation_Models_for_3D_Humans_Workshop
This is an author-produced version of the published paper. Uploaded in accordance with the University’s Research Publications and Open Access policy.
Accepted author manuscript, 1.25 MBLicence: CC BY

Cite this

@inproceedings{48341f40ecb14f52b199ecbf569cad29,

title = "N Heads Are Better Than One: Exploring Theoretical Performance Bounds of 3D Face Reconstruction Methods",

abstract = "We introduce ``N Heads Are Better Than One'', a novel approach for evaluating combinations of existing 3D face reconstruction methods. By calculating lower theoretical error bounds for method combinations on the NoW benchmark, we establish a robust set of new baselines for the task of 3D face reconstruction. Our work also provides a framework for assessing the potential of these aggregate `pseudo-foundation models,' which leverage strengths from multiple existing approaches. In doing so, we improve understanding of the performance of current methods and set targets for future foundation models to beat.",

keywords = "3D Face Reconstruction; Foundation Models; Performance Evaluation",

author = "William Rowan and Patrik Huber and Pears, {N. E.} and Andrew Keeling",

note = " This is an author-produced version of the published paper. Uploaded in accordance with the University{\textquoteright}s Research Publications and Open Access policy.",

year = "2024",

month = oct,

day = "4",

language = "English",

booktitle = "European Conference on Computer Vision Workshop (ECCVw) 2024",

publisher = "Springer Science + Business Media",

address = "United States",

}

N Heads Are Better Than One: Exploring Theoretical Performance Bounds of 3D Face Reconstruction Methods. / Rowan, William ; Huber, Patrik ; Pears, N. E. et al.
European Conference on Computer Vision Workshop (ECCVw) 2024: Foundation Models for 3D Humans. Springer Science + Business Media, 2024.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - N Heads Are Better Than One

T2 - Exploring Theoretical Performance Bounds of 3D Face Reconstruction Methods

AU - Rowan, William

AU - Huber, Patrik

AU - Pears, N. E.

AU - Keeling, Andrew

N1 - This is an author-produced version of the published paper. Uploaded in accordance with the University’s Research Publications and Open Access policy.

PY - 2024/10/4

Y1 - 2024/10/4

N2 - We introduce ``N Heads Are Better Than One'', a novel approach for evaluating combinations of existing 3D face reconstruction methods. By calculating lower theoretical error bounds for method combinations on the NoW benchmark, we establish a robust set of new baselines for the task of 3D face reconstruction. Our work also provides a framework for assessing the potential of these aggregate `pseudo-foundation models,' which leverage strengths from multiple existing approaches. In doing so, we improve understanding of the performance of current methods and set targets for future foundation models to beat.

AB - We introduce ``N Heads Are Better Than One'', a novel approach for evaluating combinations of existing 3D face reconstruction methods. By calculating lower theoretical error bounds for method combinations on the NoW benchmark, we establish a robust set of new baselines for the task of 3D face reconstruction. Our work also provides a framework for assessing the potential of these aggregate `pseudo-foundation models,' which leverage strengths from multiple existing approaches. In doing so, we improve understanding of the performance of current methods and set targets for future foundation models to beat.

KW - 3D Face Reconstruction; Foundation Models; Performance Evaluation

M3 - Conference contribution

BT - European Conference on Computer Vision Workshop (ECCVw) 2024

PB - Springer Science + Business Media

ER -