SHREC 2009 - Shape Retrieval Contest of Partial 3D Models

Call For Participation:

SHREC 2009 - Shape Retrieval Contest of partial 3D models

Introduction

There are two objectives of this shape retrieval contest:

• a) To evaluate partial similarity between query and target objects and retrieve complete 3D models that are relevant to a partial query object.
• b) To retrieve 3D models those are relevant to a query depth map. This task corresponds to a real life scenario where the query is a 3D range scan of an object acquired from an arbitrary view direction. The algorithm should retrieve the relevant 3D objects from a database.

Task description

In response to a given set of queries, the task is to evaluate similarity scores with the target models and return an ordered ranked list along with the similarity scores for each query. The set of queries either consists of partial 3D models or of range images. The participants may present ranked lists for either of the query sets or both. There is no obligation to submit rank lists for both of the query sets.

Data set

The first query set consists of 20 3D partial models which are obtained by cutting parts from complete models. The objective is to retrieve the models which the query part may belong to. The file format to represent the partial query models is the ASCII Object File Format (*.off).

This second query set is composed of 20 range images, which are acquired by capturing range data of 20 models from arbitrary view directions. The range images are captured using a desktop 3D scanner. The file format is in the ASCII Object File Format (*.off) representing the scan in a triangular mesh.

The target database is the same for both of the query sets and it contains 720 complete 3D models, which are categorized into 40 classes. In each class there are 18 models. The file format to represent the 3D models is the ASCII Object File Format (*.off).

Evaluation Methodology

We will employ the following evaluation measures: Precision-Recall curve; Average Precision (AP) and Mean Average Precision (MAP); E-Measure; Discounted Cumulative Gain; Nearest Neighbor, First-Tier (Tier1) and Second-Tier (Tier2).

Procedure

The following list is a step-by-step description of the activities:

• The participants must register by sending a message to %20shrec [at] nist.gov (SHREC[at]nist[dot]gov). Early registration is encouraged, so that we get an impression of the number of participants at an early stage.
• The database will be made available via this website. Similarly the query set will be made available, against which the evaluation will be run.
• Participants will submit the ranked lists and similarity scores for one of the two query sets or both. The participants do not have to submit query results for both of the sets. Up to 5 ranked lists per query set may be submitted, resulting from different runs. Each run may be a different algorithm, or a different parameter setting. More information on the rank list file format.
• The evaluations will be done automatically.
• The organization will release the evaluation scores of all the runs.
• The participants write a short paper describing their method and commenting the evaluation results.
• The track results are combined into a joint paper, published in the proceedings of the Eurographics Workshop on 3D Object Retrieval.
• The description of the tracks and their results are presented at the Eurographics Workshop on 3D Object Retrieval (March 29, 2009).

Schedule