Comparison of DGCNN and Point Transformer Model Performance

5.5 KB Excel file compares a convolution-based model (DGCNN) and a transformer-based model (Point Transformer). The dataset, authored by Shuhao Fu and updated in March 2026, contains results from ablation simulations and experiments on 3D shape recognition. Tags indicate analysis of human performance, local geometric structures, and sparse visual information.

Use Cases

• Analyze 'Compared Human Performance' metrics from Experiment 2 and Experiment 3 to evaluate model effectiveness.
• Use 'Ablation Simulations Revealed' data to understand the impact of 'Progressive Downsampling Operations' on model stability.
• Compare 'Local Geometric Structures' processing between DGCNN and Point Transformer architectures for 3D point clouds.
• Assess model performance on '3d shapes depicted' under conditions of 'Sparse Visual Information'.

Strengths

• Dataset is recent, last updated March 2026.
• License is permissive CC BY 4.0.
• Focuses on a direct comparison between two distinct deep learning architectures for 3D data.

Limitations

• Extremely small scale at only 5.5 KB, indicating very limited data content.
• Unknown row and column structure, as sample data and columns are unavailable.
• Scope is narrowly focused on a specific model comparison, limiting general applicability.

Provenance

Source

figshare, author Shuhao Fu.

Collection Method

Likely results from academic experiments or simulations; method not detailed.

Time Range

null

Freshness

Last updated 2026-03-13.

Geography

null