SHREC'20 track: Retrieval and classification of surface patches with similar geometric relief

THE CONTEST IS OVER! You can check out the results of our work here (link to the Computers and Graphics paper) . Moreover, you can download the dataset and ground truth of the contest from this link. Please read the licence and readme files carefully before using the dataset in your work.

The aim of this SHREC'20 track is to evaluate the performance of automatic algorithms for the retrieval and classification of relief patterns on the surface of 3D models. The goal is to estimate the similarity between two objects based only on the patterns on their surfaces, without considering their global shape. The track is organized by IMATI CNR in collaboration with the University of Verona, Italy.

Dataset, ground truth and evaluation

The dataset consists of 220 surfaces characterized by different relief patterns. Each model is partially or fully decorated with a single pattern. All models are oriented, connected and locally regular, triangle meshes. Starting from a set of "null" triangle meshes, a set of relief patterns has been applied to the meshes using a semi-automatic procedure. The models in the dataset are randomly rotated and shuffled.

Each model will be used in turn as a query against the remaining part of the database. For a given query, the goal is to retrieve the most similar objects. The relevance or non-relevance of the models for a given query, i.e. the ground truth, is established a priori, on the basis of the pattern applied to that mesh. The performance of the retrieval algorithms will be evaluated using the Precision-Recall curves, Nearest Neighbor, First Tier, Second Tier, Normalized Discounted Cumulated Gain and Average Dynamic Recall. As for the classification performances, we will adopt the percentage of success for the first retrieved item, the confusion matrix and tier images among the classes.

Registration and instructions for competitors
Each participant is requested to register to the track by sending an email to Elia Moscoso Thompson (email: elia.moscoso@ge.imati.cnr.it) with the subject SHREC'20 track: Retrieval and classification of surface patches with similar geometric reliefs. Then, an answer will be sent to each participant with further instructions on how to download the models.

Further information

• The trangulations in the dataset are in ply format.
• The results must be presented shaped as a dissimilarity matrix of the dataset.
• Each participant can send 1 to 3 runs for the algorithm(s) they used, together with the program that generated the results (binary code, including appropriate user’s instructions if needed) and a summary of the proposed method.
• The output format we expect for each run is a plain ASCII file containing a dissimilarity matrix. Entry (i,j) of the matrix is the dissimilarity measure value between the model i and the model j (i.e., the bigger the value of (i,j) is, the more different the models i and j are). We expect one file for each run.
• The evaluations will be done automatically.
• The description of the track and its results will be presented at the 13th Workshop On 3D Object Retrieval 2020 (September 4-5, 2020).