In real-world scenarios, a major limitation for shape-matching datasets is represented by having all the meshes of the same subject share their connectivity across different poses. Specifically, similar connectivities could provide a significant bias for shape-matching algorithms, simplifying the matching process and potentially leading to correspondences based on recurring triangle patterns rather than geometric correspondences between mesh parts. As a consequence, the resulting correspondence may be meaningless, and the evaluation of the algorithm may be misled.
To overcome this limitation, we introduce TACO, a new dataset where meshes representing the same subject in different poses do not share the same connectivity, and we compute new ground truth correspondences between shapes. We extensively evaluate our dataset to ensure that ground truth isometries are properly preserved. We also use our dataset to validate state-of-the-art shape-matching algorithms, verifying a degradation in performance when the connectivity gets altered.

Dataset structure

• offs: a directory containing all the triangular meshes in the dataset in OFF file format
• pairs.txt: a list of all the 420 possible pairs of shapes in the dataset
• gt_matches: a directory containing all the ground truth correspondences listed in `pairs.txt` and stored in MAT file format