text
stringlengths
36
36
fce858fe-d30b-4a72-822d-79c92d29971e
abf00ef3-e9f2-491a-9692-114c4885298c
09ca3310-232c-4f3f-b201-3632ea8c6b75
ab4b4145-563f-42f9-bfb6-97b48bdaa2c7
35823fca-da54-49a2-8d4c-365046031e2c
8e5c9c98-c1fa-4248-b473-06484a4416bc
d1563ead-9c4a-443d-8f28-515100c0b909
847fbeab-dd68-4852-95f3-3a09f754f3b7
41e70872-13dc-45e3-82b4-4cbe58d952da
af02982c-8ae9-424c-9f35-6f18763a49e1
b880f349-0ef2-4470-b597-adbd239ea6e2
6d571173-653c-400f-99cc-14111f8657aa
2336980a-6599-4f18-8a4c-6f2aa92ed536
62d59872-7198-4df5-a1e8-56b1136f6952
d2a578c6-c2e8-4a03-a809-ecb29e0c8859
1c504968-4f68-48dc-bee3-159fa197af65
38bb5f61-707d-4cf3-b1d1-a8260b7dfadc
8c6578a1-4f98-4bdb-861e-64b321b32c9f
7d33f6b7-46c0-4087-97e5-100c369ffe45
1856e7d9-d45a-4909-9206-a3d6921ec6bd
c0b7d130-2004-4450-ad94-ea3167bd9fab
c7f81b64-f398-4b75-ad96-de62f2a31ec6
66587329-5714-4561-922e-95e8b888de2e
c112d4da-5120-47d7-9b05-a4ef68baa226
ad14de7c-ea6f-4811-a590-1ee8a65275e4
587eca14-1115-4a97-aff1-664de9e3415f
66fbbbae-c740-40d1-ab71-1f882453f2c5
819523fa-f601-4ed3-aa1f-b1df5f0482bf
ff8cba46-32f2-41fc-a649-50ef32fa245c
49944f64-4fc9-44b8-bba2-7d6720c3186f
bde0f2f1-a83e-4c29-b182-502ce6d94e21
bd58f048-87aa-4b40-97ee-17bdb2dd947e
a9902d89-3aab-4bce-aa07-720cf12523f6
fb626925-969e-48ea-aeda-ef99300fb8f7
46e0152d-45be-46a4-a44a-c1967f89d3aa
5b4fcf58-6bdc-41e4-b7e0-739abc35e7e8
eac4b49b-cf4f-4925-bdbe-5ab98d0bae74
658a6c36-f169-4d0c-9811-cbbc76393e07
5d4abe32-590d-470d-8fd1-4585833afaad
6fee2c75-fded-4e56-850f-0c29c809c1ee
e102ad52-d3a9-405b-a5a2-9a77a1a9922d
952ca52f-91d7-40a8-b254-affcc831112a
b24dbfff-1a6f-4649-b60b-35ae8c7aec80
5df0a1b2-0240-4ddb-a6a8-2c48c80d7427
4af687ac-9602-48b9-acaf-c0085bc2fe53
1c9a978e-3921-44e5-8eab-f9e5401822cf
5e4d0439-eaea-463b-963b-e6ae2459f874
d2e07def-3ea4-4c31-b20c-5f5a7ed52fb9
7130970c-73d9-413d-964e-167696280422
5dc686b7-d094-426b-8a79-7018ce1f4d0f
60a5958c-4530-4eba-886a-a3474ab1fc7a
084fe4d0-b55d-4de4-a26d-32d7735e9621
5b4cc5cb-22f0-4814-aa24-8d51ba50000a
60475f06-2613-4496-939a-5381e7807349
16b05645-80cd-45e6-ad95-c9a9cc035d6d
83b377ef-967a-4380-b3ac-321dbf79f72a
3cf6545b-6476-41f4-a526-6d1da55588e3
c8749317-8f4c-443a-9f82-6b18edbcac23
39feb026-bcf8-4a61-89ad-9f5566c2c7bf
51fc36b3-e769-4617-b087-3826b280cad3
a2fc6bb5-ad56-460e-a8ae-64db8080d5ae
ef31e3a5-a004-4fc8-8d05-8475afa2602a
43d60820-1daa-472f-be4d-6802746df7d8
3059469a-03fc-4ae0-bbf7-b08187d1b290
823affad-e570-4091-a85f-da7aed524500
b22e7d54-2198-4ff8-9ba4-bf444bf4a787
1215d140-b34c-43f3-82dc-c3f102bbf1e4
6bc7a29f-3397-4549-9ee2-d98fa93da873
206e9042-02a3-48ac-bed0-45861fe658cb
37b18db8-a85a-49dd-a594-68cb1dd5ed0e
c529c120-c1fb-4504-97c7-52912c2aa3c0
10114022-4de6-43cd-af19-b3bc5cbeae31
3b6e9197-fbe8-48cf-891f-16ca9ebbc6aa
eba56e4f-7ec8-4d47-9380-e69928323e94
fa0e1a71-2174-43ef-a694-a4af360573a4
d6817560-492c-4de8-93e9-56df0c3b4ddc
e68d09a0-fe18-483c-8a7d-24b32f9baa9e
44e3a281-953f-439b-aa58-01a5a362a31f
f1c3d2f5-e6bf-4899-a40d-12489d1ebb11
c4dbfed6-f922-4213-a574-da37c7afd61c
99c10780-e9a9-431a-af53-fda0b4fae201
c95ba288-2cca-46f2-b606-f4e4d145d40e
11b8d7e3-9c5e-4652-84ef-91504a07f332
66a58e88-a8a3-49c8-94db-949274cb5be0
831704c0-37fc-47aa-9ac0-3681ee652690
f7b3e85b-7681-48b3-97cb-6b0a5705022e
515ec288-a236-4a94-9b3d-449667088c82
83d9f439-642d-47b8-bc95-1a9442a0f4f6
dcf60f14-5b6e-4550-b1ed-0befd0849074
e02bcacc-92af-4fe0-8c57-e42ff3f74dcf
1377b49e-4a24-4e91-9fdb-709911b72f9a
76a7ea80-0a81-4b4f-983e-a7e36fa26685
e3b4afb7-d65c-4185-a038-9b9235caf342
2558806a-cf6c-42dd-8323-ab80e7b7dc75
9bc33576-bcb6-42a5-b040-3220456f268f
410ccf87-4217-4d87-836b-4af20b355d87
0efd9fc1-0f49-403d-9252-93f119a3ea3c
719fe670-72a7-4da9-be7d-0bd2afb742ae
cc27440d-408b-4280-9436-e844ad5e7d6f
62bbef54-0ba2-4990-a9e3-3dea114fa9cd

Dataset Preparation.

Our Ego2Exo benchmark is curated from the EgoExo4D dataset released by Meta. A complete documentation of the dataset preparation including selecting video segments, narrations and label set is available at this colab notebook. Note that you have to first download the annotations and keystep labels from the original dataset.

Dataset Download Instructions

The original video segments should be downloaded from the original EgoExo4D website from Meta by following their guidelines. You might have to first sign a license form to access and download the dataset, which might take upto 48 hours for getting approved. You can download the cli installer to enable faster downloads. Make sure that the installation works by running egoexo --help which should return the documentation guide.

In our paper, we only used the pre-extracted Omnivore features for all the video segments as input to the classification network. If you only want to download the pre-extracted features for the takes corresponding to our dataset, use the following command. Make sure to provide the argument data_dir or specify where the files should be downloaded.

sh omnivore_features_download.sh <data_dir>

You can also download the complete videos corresponding to our takes. The whole dataset is quite large, so it is recommended that you only download the videos used in our benchmark with the following command. Make sure to provide the argument data_dir or specify where the files should be downloaded.

sh takes_download.sh <data_dir>
Downloads last month
47
Edit dataset card