Tencent Dynamic Colored Mesh Dataset (TDMD)

As a typical representation of three-dimensional (3D) graphics, 3D dynamic colored mesh (DCM) plays a crucial role in various applications such as animated advertisement, digital entertainment, and education. These applications generally require high-quality meshes to provide a better quality of experience (QoE) through improved immersion. Thus, accurately evaluating the influence of typical distortions on mesh quality is an essential task.

To study the influence of typical distortions on DCM perception and contribute to the development of objective metrics, we construct a DCM subjective database called Tencent DCM (TDMD). The Tencent dynamic colored mesh database (TDMD) is proposed. The database contains eight reference DCM objects with six typical distortions, including color noise (CN), geometrical Gaussian noise (GN), texture map downsampling (DS), mesh decimation (MD), MPEG lossy compression (MLC), and texture map compression (TC). Each type of distortion has 5 to 7 degrees. The DCMs were converted into processed video sequences (PVS) and a large-scale subjective experiment was conducted. In total, 303 distorted DCM samples with mean opinion scores (MOS) were provided. As far as we know, this is the largest DCM subjective database publicly available.