Trajectory-to-Video Framework for Transportation Mode Recognition

A 168.9 MB repository contains core code for a dual-path framework converting trajectory data into video representations for multiclass transportation mode recognition. The work, authored by Kunkun Fan and shared on figshare, was last updated in April 2026.

Use Cases

• Implement trajectory-to-video conversion for feature extraction in mode recognition models.
• Benchmark the dual-path framework against other video-based classification methods.
• Reproduce experiments for multiclass recognition of transportation modes like walking, cycling, or driving.

Strengths

• Includes core implementation code for a published framework.
• Released under a permissive CC-BY-4.0 license.
• Updated in 2026 indicating recent maintenance.

Limitations

• No actual dataset or sample data is provided; only code.
• Unknown row count or volume of training/validation data.
• Performance depends on external trajectory data sources.

Provenance

Source

Kunkun Fan via figshare.

Collection Method

Code implementation for a research framework.

Freshness

Last updated April 2026.