The 3rd SHApe Recovery from Partial textured 3D scans (SHARP) Workshop and Challenge will be held in conjunction with CVPR on June 19, 2022 (TBC). Research on data-driven 3D reconstruction of shapes has been very active in the past years thanks to the availability of large datasets of 3D models. However, current methods did not focus enough on two important aspects: (1) reconstructing full 3D objects with shape and texture at the same time, (2) reconstructing sharp edges that tend to be smoothed by 3D scanning.  The goal of this workshop is to promote the development of methods to recover a complete 3D scan from its partial acquisition. This should in turn foster the development of 3D modelling and processing techniques that exploit both geometry and texture.

This workshop will host a paper submission track and a competition:

  • The paper submission track encourages participants to submit novel contributions on data-driven shape and texture processing. A call for papers specifying the topics of interest can be found below.
  • The competition focuses on the reconstruction of full high-resolution 3D meshes from partial or noisy 3D scans and includes 2 challenges and 3 datasets:
    • The first challenge consists of the recovery of textured 3D scans from partial acquisition. It involoves 2 tracks:
      • Track 1: Recovering textured human body scans from partial acquisitions. The dataset used in this scope is the 3DBodyTex.v2 dataset, containing 2500 textured 3D scans. It is an extended version of the original 3DBodyTex.v1 dataset, first published in the 2018 International Conference on 3D Computer Vision, 3DV 2018.
      • Track 2: Recovering textured object scans from partial acquisitions. It involves the recovery of generic object scans from the 3DObjTex.v1 dataset, which is a subset from the ViewShape online repository of 3D scans. This dataset contains over 2000 various generic objects with different levels of complexity in texture and geometry.
    • The second challenge focuses on the recovery of fine object details in the form of sharp edges from noisy sparse scans with smooth edges. The CC3D-PSE dataset which is a new version of the CC3D dataset, introduced at the 2020 IEEE International Conference on Image Processing (ICIP), will be used for this purpose. It contains over 50k pairs of CAD models and their corresponding 3D scans. Each pair of scan and CAD model is annotated with parametric sharp edges. Given a 3D scan with smooth edges, the goal is to reconstruct the corresponding CAD model as a triangular mesh, with sharp edges approximating the ground-truth sharp edges. The second challenge invovles 2 tracks:
      • Track 1: Recovering linear sharp edges. A subset of the CC3D-PSE dataset is considered in this track which includes only linear sharp edges.
      • Track 2:  Recovering sharp edges as linear, circular, and spline segments. The whole CC3D-PSE will be used in this track.

Sponsor

 

 

Call for Participation (Challenges)

We propose two challenges. The task of the challenges is to reconstruct a full 3D textured mesh from a partial 3D scan. The first challenge focuses on the recovery of textured 3D scans from partial acquisition, while the second consists of the recovery of fine object details in the form of sharp edges from noisy sparse scans with smooth edges.

🏆 An overall 8k€ will be awarded as cash prizes to the winners.

Challenge 1

Textured Partial Scan Completion

Given Partial Textured Scans, the goal is to recover scans with completed shape and texture. More information can be found here.

 

Challenge 2

Sharp Edge Recovery in Object Scans

Given an object scan, the goal is recover sharp edges and the geometry of the corresponding CAD model. More information can be found here.

Track 1

Recovery of Partial Textured Human Body Scans

Track 2

Recovery of Partial Textured Object Scans

Track 1

Recovery of Linear Sharp Edges

Track 2

Recovery of Generic Sharp Edges

Call for Papers

The main focus of SHARP is to encourage paper submissions on high-resolution 3D shape and texture recovery from partial data, especially as accompanying papers to the challenge submissions. In addition, all topics that relate to and serve the goal of data-driven shape and texture processing are of interest. This includes original contributions at different levels of data processing; for different industrial applications, as well as proposals for new evaluation metrics and relevant original datasets. Topics of interest include, but are not limited to:

  • Textured 3D data representation and evaluation
  • Textured 3D scan feature extraction
  • Generative modelling of textured 3D scans
  • Learning-based 3D reconstruction
  • Joint texture and shape matching
  • Joint texture and shape completion
  • Semantic 3D data reconstruction
  • Effective 3D and 2D data fusion
  • Textured 3D data refinement
  • 3D feature edge detection and refinement
  • High-level representations of 3D data
  • CAD modeling from unstructured 3D data

All accepted papers will be included in the CVPR 2022 conference proceedings. The papers will be peer-reviewed and they must comply with the CVPR 2022 proceedings style and format. More details about the submission formats can be found in the submission page.

Important Dates

Challenges

Website opened: 15th January 2022

Release of training datasets: 20th January 2022

Registration deadline: 15th March 2022 (tbc)

Release of evaluation datasets: 15th April 2022 (tbc)

Submission of results: 20th May 2022 (tbc)

Paper Submission

Website opened: 15th January 2022

Paper submission deadline:  1st May 2022 (tbc)

Final decision to authors: 15th May 2022 (tbc)

Camera-ready submission deadline: 22nd May 2022 (tbc)

Organizers

DjamilaAouadaFeb20_v1
Djamila Aouada

Chair

SnT, University of Luxembourg

djamila.aouada@uni.lu

IMG_3274_ne
Kseniya Cherenkova

Co-Chair

Artec3D, SnT

kcherenkova@artec-group.com

Anis_Kacem2
Anis Kacem

 

SnT, University of Luxembourg

anis.kacem@uni.lu

azizal
Sk Aziz Ali

 

SnT, University of Luxembourg

skaziz.ali@uni.lu

Elona_Dupont
Elona Dupont

 

SnT, University of Luxembourg

elona.dupont@uni.lu

gleb_
Gleb Gusev

 

Artec3D

gleb@artec-group.com

Ottersten2
Bjorn Ottersten

 

SnT, University of Luxembourg

bjorn.ottersten@uni.lu

DA457009-9004-43AA-A72C-35D40E8FC8F7_1_105_c
David Fofi

 

University of Burgundy

david.fofi@u-bourgogne.fr

TEST

A dataset containing 400 real, high-resolution human scans of 200 subjects (100 males and 100 females in two poses each) with high-quality texture and their corresponding low-resolution meshes, with automatically computed ground-truth correspondences. See the following table.

TEST

A dataset containing 400 real, high-resolution human scans of 200 subjects (100 males and 100 females in two poses each) with high-quality texture and their corresponding low-resolution meshes, with automatically computed ground-truth correspondences. See the following table.

TEST

A dataset containing 400 real, high-resolution human scans of 200 subjects (100 males and 100 females in two poses each) with high-quality texture and their corresponding low-resolution meshes, with automatically computed ground-truth correspondences. See the following table.

TEST

A dataset containing 400 real, high-resolution human scans of 200 subjects (100 males and 100 females in two poses each) with high-quality texture and their corresponding low-resolution meshes, with automatically computed ground-truth correspondences. See the following table.