GTEA Benchmark (Action Segmentation)

28 rows 20 with code listed 5 metrics Dataset page

Action Segmentation is a challenging problem in high-level video understanding. In its simplest form, Action Segmentation aims to segment a temporally untrimmed video by time and label each segmented part with one of pre-defined action labels. The results of Action Segmentation can be further used as input to various applications, such as video-to-text and action localization.

Source: TricorNet: A Hybrid Temporal Convolutional and Recurrent Network for Video Action Segmentation

The archive carries no text for this table; the description above is the archive's text for the task Action Segmentation. archive 2025-07-28

Over time archive 2025-07-28

Metric Extra training data

The chart needs JavaScript; the table below carries every value.

Direction inferred from the metric name, not from the archive: F1@50% (higher is better), F1@25% (higher is better), F1@10% (higher is better), Acc (higher is better). Not inferred (points only, no best-so-far line): Edit. Points are placed at the row's paper date; 27 of 28 rows carry one.

Results archive 2025-07-28

Archive rows end at the archive snapshot, 2025-07-28: no result published after that date is in this table. Rank is the archive's row order at that snapshot; not re-ranked here. Metric values are the archive's strings. Column headers sort the table in your browser; each row keeps its archive rank.

								Paper	Code		Ran Syntology	Report
1	Semantic2Graph	91.3	94.2	95.7	89.8	92.0	✓	Paper	–	2022	no code linked	report
2	FACT	87.5	95.6	96.1	84.5	93.5	–	Paper	Code	2024	linked, not harvested	report
3	DiffAct	84.7	91.5	92.5	82.2	89.6	–	Paper	Code	2023	5 of 9 ran · 4 unverified	report
4	BaFormer	83.5	91.3	92.0	83.0	88.7	–	Paper	Code	2024	linked, not harvested	report
5	SF-TMN(ASFormer)	83.1	90.7	91.9	83.0	88.9	–	Paper	–	2023	no code linked	report
6	Br-Prompt+ASFormer	83.0	92.0	94.1	81.2	91.6	✓	Paper	Code	2022	2 of 2 ran · 0 unverified	report
7	DPRN	82.9	92.0	92.9	82.0	90.9	–	Paper	–	2022	no code linked	report
8	BIT	82.6	92.8	94.8	82.0	92.6	–	Paper	–	2023	no code linked	report
9	CETNet	81.3	91.2	91.8	80.3	87.9	–	Paper	Code	2022	linked, not harvested	report
10	UVAST	81	91.3	92.7	80.2	92.1	–	Paper	Code	2022	2 of 2 ran · 0 unverified	report
11	ASRF	79.8	87.8	89.4	77.3	83.7	–	Paper	Code	2020	0 of 11 ran · 11 unverified	report
12	ASFormer	79.2	88.8	90.1	79.7	84.6	–	Paper	Code	2021	1 of 1 ran · 0 unverified	report
13	SSTDA	78.0	89.1	90.0	79.8	86.2	–	Paper	Code	2020	3 of 3 ran · 0 unverified	report
14	ETSN	77.9	90.0	91.1	78.2	86.2	–	Paper	Code	2021	linked, not harvested	report
15	C2F-TCN	77.7	88.8	90.3	80.8	86.4	–	Paper	Code	2021	linked, not harvested	report
16	BCN	77.3	87.1	88.5	79.8	84.4	–	Paper	Code		linked, not harvested	report
17	SSTDA + HASR	76.4	88.6	90.9	78.7	87.5	–	Paper	Code	2021	linked, not harvested	report
18	DA	76.2	88.4	90.5	80.0	85.8	–	Paper	–	2021	no code linked	report
19	MS-TCN++	76.0	85.7	88.8	80.1	83.5	–	Paper	Code	2020	linked, not harvested	report
20	MS-TCN++(sh)	75.9	86.2	88.2	79.7	83.0	–	Paper	Code	2020	linked, not harvested	report
21	ASRF + HASR	74.8	87.2	89.2	76.9	84.5	–	Paper	Code	2021	linked, not harvested	report
22	MS-TCN	74.6	85.4	87.5	79.2	81.4	–	Paper	Code	2019	linked, not harvested	report
23	EUT	74	87.2	88.2	77	83.9	–	Paper	Code	2022	linked, not harvested	report
24	DS-TCN	72.84	85.44	88.30	78.10	84.05	–	Paper	–	2021	no code linked	report
25	AUL	67.3	85.5	88.2	69.2	84.0	–	Paper	Code	2023	linked, not harvested	report
26	TDRN	62.7	74.4	79.2	70.1	74.1	–	Paper	–	2018	no code linked	report
27	ED-TCN	56.0	69.3	72.2	64.0	-	–	Paper	Code	2016	2 of 21 ran · 19 unverified	report
28	ST-CNN	41.9	54.4	58.7	60.6	-	–	Paper	–	2016	no code linked	report

All 28 rows shown. 28 link to a paper page on this site; 2 are marked as using additional training data in the archive. No GitHub stars are tracked; "Code" is the first repository the archive lists for the row. The archive carries no row tags, review links or community-submitted rows for this table; none are shown. archive 2025-07-28

Syntology Ran reads "N of M ran · U unverified": of the M code samples Syntology harvested from repositories linked to that row's paper (joined by arXiv id), N executed on a synthesized input and the other U = M−N are unverified (harvested, no recorded run). It counts code from repositories linked to that row's paper, not this result: the row's number was not reproduced and nothing here is a correctness claim. The other cell texts mean no graph line for the row: "linked, not harvested" (the archive links code, Syntology has not harvested it), "no code linked" (no code link in the archive), "not matched" (the row's paper URL matched no paper on this site). 7 rows have a graph line, from 7 distinct papers; 6 rows (6 papers) have at least one sample that ran. Counting each paper once: Syntology ran 15 of 49 samples; 34 unverified. Separately, 4 of those 49 are pointer-only (licence): the site points at that code rather than redistributing it, a licence property recorded for ran and unverified samples alike; each cell's tooltip carries the row's own pointer-only count. Read from the graph 2026-09-24. Per-sample status is on the paper page.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections