josauder/coralscapesV2

Optional configurations

The default configuration remains the original image + label dataset and is loaded with load_dataset(REPO). Optional data is stored in separate Parquet shards and is downloaded only when its configuration is requested.

Instances

load_dataset(REPO, "instances") adds a uint16 instance_ids image. Fish is class 25 and per-fish IDs are 25 * 2000 + instance_index; a binary mask is instance_ids == instance_id.

Sequences

load_dataset(REPO, "sequences", streaming=True) returns frames ordered from offset -20 through +10, with the labeled anchor at anchor_index == 20. At a genuine video boundary, unavailable frames repeat the nearest available frame and source_frame_offsets records that padding.

CoralscapesV2

Paper: CoralscapesV2: Panoptic and Fine-Grained Visual Scene Understanding in Coral Reefs (ECCV 2026 Workshop on Marine Vision) · Project page: josauder.github.io/coralscapesv2

CoralscapesV2 is a general-purpose scene-understanding dataset for coral reefs and the first dataset for panoptic segmentation in reefs. It extends the original Coralscapes dataset along every measurable axis: more annotated frames, higher annotation density and quality, a refined label set of 95 fine-grained visual categories (up from 39), and 65,738 exhaustive fish instance masks annotated using video context.

Dataset Structure

CoralscapesV2 spans 2433 densely-annotated video frames at 1024×2048px resolution gathered from 41 dive sites in the Red Sea region, labeled in a consistent and speculation-free manner over 95 fine-grained visual categories.

The released semantic label masks (label) use the 95-class fine-grained taxonomy: class ids are given by id2label.json and colors by label2color.json. Every fine class also maps deterministically onto one of the 39 legacy Coralscapes classes (rightmost column of the table below), so models and evaluations targeting the original 39-class taxonomy remain fully supported. Per-fish instance masks are available through the optional instances configuration (see above).

Data Splits

The dataset provides three splits

• train - 1790 images across 34 dive sites
• validation - 170 images across 3 dive sites
• test - 473 images across 5 dive sites

Fine-Grained Visual Categories

CoralscapesV2 refines the original 39-class label set into 95 fine-grained classes. The newly added classes fall into three groups:

• More taxa & morphotypes. Clearly identifiable hard corals (lobophylliidae, porites, fungiidae, pavona, galaxea, goniopora, seriatopora) and soft corals (soft coral (malacalcyonacea), nephtheidae, xeniidae) are separated out. For distinct morphologies whose taxon cannot be identified confidently, morphologic classes brain coral and thin plate/encrusting coral are used, and catch-all hard coral/soft coral classes guarantee full coverage while preserving the hard–soft distinction. The black coral cirrhipathes and the sabellid feather worm are also included.
• Composite health classes. Taxonomic classes are split into alive / bleached / dead / algae covered states where visually distinguishable.
• Auxiliary classes. human-made structure captures permanent underwater infrastructure (coral nursery tables, concrete blocks, mooring ropes), and the legacy background is disambiguated into water column (an unobstructed view into the water) and water surface (light ripples on surface waves).

The full 95-class label set with colors and the mapping to the 39 legacy classes:

Id	Name	Color	RGB	Legacy (39-class)
1	acanthaster planci	#B3F5EA	179, 245, 234	crown of thorn
2	acropora alive	#EC80FF	236, 128, 255	acropora alive
3	acropora bleached	#FFECFE	255, 236, 254	branching bleached
4	acropora dead	#763F80	118, 63, 128	branching dead
5	algae covered brain coral	#D0CF3A	208, 207, 58	algae covered substrate
6	algae covered branching coral	#0647F5	6, 71, 245	algae covered substrate
7	algae covered lobophylliidae	#32D082	50, 208, 130	algae covered substrate
8	algae covered massive coral	#43B6D0	67, 182, 208	algae covered substrate
9	algae covered pocillopora	#E16D4D	225, 109, 77	algae covered substrate
10	algae covered porites	#AD5182	173, 81, 130	algae covered substrate
11	algae covered substrate	#7DA37D	125, 163, 125	algae covered substrate
12	algae covered table acropora	#30A2DD	48, 162, 221	algae covered substrate
13	anemone	#00FFBD	0, 255, 189	anemone
14	background	#1DA2D8	29, 162, 216	background
15	brain coral alive	#724E78	114, 78, 120	meandering alive
16	brain coral bleached	#DB43EE	219, 67, 238	meandering bleached
17	brain coral dead	#4EAC63	78, 172, 99	meandering dead
18	branching coral alive	#E25B9D	226, 91, 157	branching alive
19	branching coral bleached	#FCE7F0	252, 231, 240	branching bleached
20	branching coral dead	#7B3256	123, 50, 86	branching dead
21	cirrhipathes	#06BD16	6, 189, 22	other coral alive
22	dark	#1F1F1F	31, 31, 31	dark
23	dead giant clam	#599B86	89, 155, 134	dead clam
24	feather worm	#9FBA26	159, 186, 38	other animal
25	fish	#FFFF00	255, 255, 0	fish
26	fungiidae alive	#6485F8	100, 133, 248	other coral alive
27	fungiidae bleached	#C88A1B	200, 138, 27	other coral bleached
28	fungiidae dead	#035C2C	3, 92, 44	other coral dead
29	galaxea alive	#60BC6C	96, 188, 108	other coral alive
30	galaxea bleached	#43F33C	67, 243, 60	other coral bleached
31	galaxea dead	#B0E10B	176, 225, 11	massive/meandering dead
32	giant clam	#BDFFEA	189, 255, 234	clam
33	goniopora alive	#B50578	181, 5, 120	other coral alive
34	goniopora dead	#B3BF1E	179, 191, 30	other coral dead
35	hard coral alive	#6DE8BC	109, 232, 188	other coral alive
36	hard coral bleached	#46F795	70, 247, 149	other coral bleached
37	hard coral dead	#720FCB	114, 15, 203	other coral dead
38	hard substrate	#7D7D7D	125, 125, 125	hard substrate
39	human	#FF0000	255, 0, 0	human
40	human-made structure	#97AB50	151, 171, 80	hard substrate
41	leather coral alive	#56B85E	86, 184, 94	other coral alive
42	lobophylliidae alive	#C949EA	201, 73, 234	meandering alive
43	lobophylliidae bleached	#FDBB4F	253, 187, 79	meandering bleached
44	lobophylliidae dead	#396ADC	57, 106, 220	meandering dead
45	massive alive	#EC9615	236, 150, 21	massive/meandering alive
46	massive bleached	#FFF8E4	255, 248, 228	massive/meandering bleached
47	massive dead	#865612	134, 86, 18	massive/meandering dead
48	meandering coral alive	#E6C100	230, 193, 0	meandering alive
49	meandering coral bleached	#FBF3D8	251, 243, 216	meandering bleached
50	meandering coral dead	#77640E	119, 100, 14	meandering dead
51	millepora	#F49673	244, 150, 115	millepora
52	millepora bleached	#A4E234	164, 226, 52	branching bleached
53	millepora dead	#7780AE	119, 128, 174	branching dead
54	nephtheidae alive	#846D5B	132, 109, 91	other coral alive
55	other animal	#00FFFF	0, 255, 255	other animal
56	other coral alive	#E07677	224, 118, 119	other coral alive
57	other coral bleached	#FAE0E1	250, 224, 225	other coral bleached
58	other coral dead	#723C3D	114, 60, 61	other coral dead
59	pavona alive	#0DBF66	13, 191, 102	other coral alive
60	pavona bleached	#76F982	118, 249, 130	other coral bleached
61	pavona dead	#8D82AF	141, 130, 175	other coral dead
62	pocillopora alive	#FF9296	255, 146, 150	pocillopora alive
63	pocillopora bleached	#FADBDC	250, 219, 220	branching bleached
64	pocillopora dead	#824242	130, 66, 66	branching dead
65	porites alive	#FFD147	255, 209, 71	massive/meandering alive
66	porites bleached	#D50A19	213, 10, 25	massive/meandering bleached
67	porites dead	#2818F3	40, 24, 243	massive/meandering dead
68	rubble	#A19980	161, 153, 128	rubble
69	sand	#C2B280	194, 178, 128	sand
70	sea cucumber	#00E7FF	0, 231, 255	sea cucumber
71	sea urchin	#008EFF	0, 142, 255	sea urchin
72	seagrass	#7DDE7D	125, 222, 125	seagrass
73	seriatopora alive	#BF1306	191, 19, 6	branching alive
74	seriatopora bleached	#F759BF	247, 89, 191	branching bleached
75	seriatopora dead	#4B8380	75, 131, 128	branching dead
76	soft coral (malacalcyonacea) alive	#939036	147, 144, 54	other coral alive
77	soft coral (malacalcyonacea) bleached	#07E3F7	7, 227, 247	other coral bleached
78	sponge	#F05050	240, 80, 80	sponge
79	stylophora alive	#FF6FC2	255, 111, 194	stylophora alive
80	stylophora bleached	#FFDCF1	255, 220, 241	branching bleached
81	stylophora dead	#7B365E	123, 54, 94	branching dead
82	table acropora alive	#BD77FF	189, 119, 255	table acropora alive
83	table acropora bleached	#F4E9FD	244, 233, 253	branching bleached
84	table acropora dead	#553574	85, 53, 116	table acropora dead
85	thin plate/encrusting coral alive	#5B1CBE	91, 28, 190	other coral alive
86	thin plate/encrusting coral bleached	#2AD139	42, 209, 57	other coral bleached
87	thin plate/encrusting coral dead	#0813F2	8, 19, 242	other coral dead
88	transect line	#00FF00	0, 255, 0	transect line
89	transect tools	#08CD0C	8, 205, 12	transect tools
90	trash	#FF0086	255, 0, 134	trash
91	turbinaria	#E4FF77	228, 255, 119	turbinaria
92	turbinaria dead	#A91E46	169, 30, 70	other coral dead
93	water column	#1F00A6	31, 0, 166	background
94	water surface	#00FFF8	0, 255, 248	background
95	xeniidae alive	#8B8EA8	139, 142, 168	other coral alive

CoralscapesV2 extends the class set from 39 (a) to 95 (b) fine-grained classes, shown here alongside their per-split polygon (c) and pixel (d) counts.

License

This dataset is licensed under the Apache License 2.0.

Citation

If you find this dataset useful, please cite CoralscapesV2:

@inproceedings{sauder2026coralscapesv2,
  title={CoralscapesV2: Panoptic and Fine-Grained Visual Scene Understanding in Coral Reefs},
  author={Sauder, Jonathan and Ruckli, Thomas and Strodomskyt{\.e}, Gabriel{\.e} and Abdallah, Ibrahim Souleiman and Abdi, Rahma Hassan and Awaleh, Djama Goumaneh and Farah, Mohamed Houssein and Nour, Moustapha and Saad, Osama Sharhubil and Altaib, Mustafa Mohammed Khalafallah and Kteifan, Maysoon and Alsoqi, Farah and Zgool, Eyad and Al-Omari, Jafar and Gebreluel, Temesgen Gebremeskel and Abdulkerim, Zekaria Zekeria and Ghirmay, Meron and Beraki, Teklehaimanot and Tuia, Devis and Banc-Prandi, Guilhem},
  booktitle={Proceedings of the European Conference on Computer Vision (ECCV) Workshop on Marine Vision},
  year={2026}
}

CoralscapesV2 extends the original Coralscapes dataset:

@misc{sauder2025coralscapesdatasetsemanticscene,
        title={The Coralscapes Dataset: Semantic Scene Understanding in Coral Reefs},
        author={Jonathan Sauder and Viktor Domazetoski and Guilhem Banc-Prandi and Gabriela Perna and Anders Meibom and Devis Tuia},
        year={2025},
        eprint={2503.20000},
        archivePrefix={arXiv},
        primaryClass={cs.CV},
        url={https://arxiv.org/abs/2503.20000},
  }