Finebadminton-20K: 20,757 Annotated Hit Instances from 70 Badminton Match Videos

20,757 annotated hit instances from 70 full-match badminton videos, comprising a public subset of the FineBadminton dataset. This release contains 2,066 annotated rallies for fine-grained video understanding. The dataset was created by iLearn-Lab and last updated in April 2026.

Use Cases

• Train models for fine-grained action recognition based on annotated hit instances.
• Develop rally detection and segmentation algorithms based on annotated rally boundaries.
• Analyze player tactics and shot sequences based on the temporal sequence of annotated hits.
• Benchmark video understanding models for sports analytics using the structured video annotations.

Strengths

• 20,757 annotated hit instances provide a substantial basis for model training.
• Annotations cover 70 full-match videos, offering contextual completeness.
• Includes 2,066 annotated rallies for higher-level event analysis.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is a subset, representing approximately 62% of the full collection's annotations.

Provenance

Source

iLearn-Lab

Collection Method

Likely involves manual or semi-automated annotation of badminton match videos.

Time Range

null

Freshness

Last updated 2026-04-02 02:03:43; freshness should be verified.

Geography

null