search

Vision-CAIR / 3DCoMPaT-v2

3DCoMPaT++: An improved large-scale 3D vision dataset for compositional recognition
BSD 3-Clause "New" or "Revised" License
77 stars 7 forks source link

Are segmentation masks for different parts of each objects available? #11

Open edoardoMalaspina opened 4 days ago

edoardoMalaspina commented 4 days ago

Hi, congratulations for your work and thank you for making this dataset publicly available. I'm trying to use this dataset to train a segmentation model and since each object is divided in components i was expecting to find also the segmentation masks of each part. Is it something that i can download in some way? If not, were you able to obtain them in other ways, like starting from the 3d models? Thank you in advance!

HabibSlim commented 4 days ago

Hello Eduardo, we do have 2D semantic segmentation masks for each part. Is this what you are looking for? Or do you mean 3D mesh segments?

Best regards, Habib

On Wed, Nov 13, 2024 at 6:49 PM edoardoMalaspina @.***> wrote:

Hi, congratulations for your work and thank you for making this dataset publicly available. I'm trying to use this dataset to train a segmentation model and since each object is divided in components i was expecting to find also the segmentation masks of each part. Is it something that i can download in some way? If not, were you able to obtain them in other ways, like starting from the 3d models? Thank you in advance!

— Reply to this email directly, view it on GitHub https://urldefense.com/v3/__https://github.com/Vision-CAIR/3DCoMPaT-v2/issues/11__;!!Nmw4Hv0!xQTKcPVQlbj7rHFFbYfin5W23H6Fd5hCxTHILNVPrim-xNAjwveimVnhkIRPpCqGtwR0MfDQfuzae8FUx3ptm1Go8oih1g$, or unsubscribe https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AG5EQIODQE2P7ITUNY6OSLT2ANYHVAVCNFSM6AAAAABRWZ2PJSVHI2DSMVQWIX3LMV43ASLTON2WKOZSGY2TKOJVGM4TIMQ__;!!Nmw4Hv0!xQTKcPVQlbj7rHFFbYfin5W23H6Fd5hCxTHILNVPrim-xNAjwveimVnhkIRPpCqGtwR0MfDQfuzae8FUx3ptm1E1eVHsMQ$ . You are receiving this because you are subscribed to this thread.Message ID: @.***>

--

This message and its contents, including attachments are intended solely for the original recipient. If you are not the intended recipient or have received this message in error, please notify me immediately and delete this message from your computer system. Any unauthorized use or distribution is prohibited. Please consider the environment before printing this email.

edoardoMalaspina commented 4 days ago

Thanks for the fast reply! I think that i already found the 3D mesh segments (the various components inside the .gltf files) but i'm not able to find the 2D semantic segmentation masks, can you please explain how to get them? Are they referred to a frontal view of the rendered 3D object?

Thanks again!

Il giorno mer 13 nov 2024 alle ore 16:52 Habib Slim < @.***> ha scritto:

Hello Eduardo, we do have 2D semantic segmentation masks for each part. Is this what you are looking for? Or do you mean 3D mesh segments?

Best regards, Habib

On Wed, Nov 13, 2024 at 6:49 PM edoardoMalaspina @.***> wrote:

Hi, congratulations for your work and thank you for making this dataset publicly available. I'm trying to use this dataset to train a segmentation model and since each object is divided in components i was expecting to find also the segmentation masks of each part. Is it something that i can download in some way? If not, were you able to obtain them in other ways, like starting from the 3d models? Thank you in advance!

— Reply to this email directly, view it on GitHub < https://urldefense.com/v3/__https://github.com/Vision-CAIR/3DCoMPaT-v2/issues/11__;!!Nmw4Hv0!xQTKcPVQlbj7rHFFbYfin5W23H6Fd5hCxTHILNVPrim-xNAjwveimVnhkIRPpCqGtwR0MfDQfuzae8FUx3ptm1Go8oih1g$>,

or unsubscribe < https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AG5EQIODQE2P7ITUNY6OSLT2ANYHVAVCNFSM6AAAAABRWZ2PJSVHI2DSMVQWIX3LMV43ASLTON2WKOZSGY2TKOJVGM4TIMQ__;!!Nmw4Hv0!xQTKcPVQlbj7rHFFbYfin5W23H6Fd5hCxTHILNVPrim-xNAjwveimVnhkIRPpCqGtwR0MfDQfuzae8FUx3ptm1E1eVHsMQ$>

. You are receiving this because you are subscribed to this thread.Message ID: @.***>

--

This message and its contents, including attachments are intended solely for the original recipient. If you are not the intended recipient or have received this message in error, please notify me immediately and delete this message from your computer system. Any unauthorized use or distribution is prohibited. Please consider the environment before printing this email.

— Reply to this email directly, view it on GitHub https://github.com/Vision-CAIR/3DCoMPaT-v2/issues/11#issuecomment-2474013437, or unsubscribe https://github.com/notifications/unsubscribe-auth/AW4ZEWB6FAGOYHJUILZFV432ANYURAVCNFSM6AAAAABRWZ2PJSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINZUGAYTGNBTG4 . You are receiving this because you authored the thread.Message ID: @.***>

HabibSlim commented 4 days ago

Dear Eduardo,

For 2D masks refer to this:

https://3dcompat-dataset.org/doc/loaders-2D.html

Best, Habib

On Wed, 13 Nov 2024, 19:05 edoardoMalaspina, @.***> wrote:

Thanks for the fast reply! I think that i already found the 3D mesh segments (the various components inside the .gltf files) but i'm not able to find the 2D semantic segmentation masks, can you please explain how to get them? Are they referred to a frontal view of the rendered 3D object?

Thanks again!

Il giorno mer 13 nov 2024 alle ore 16:52 Habib Slim < @.***> ha scritto:

Hello Eduardo, we do have 2D semantic segmentation masks for each part. Is this what you are looking for? Or do you mean 3D mesh segments?

Best regards, Habib

On Wed, Nov 13, 2024 at 6:49 PM edoardoMalaspina @.***> wrote:

Hi, congratulations for your work and thank you for making this dataset publicly available. I'm trying to use this dataset to train a segmentation model and since each object is divided in components i was expecting to find also the segmentation masks of each part. Is it something that i can download in some way? If not, were you able to obtain them in other ways, like starting from the 3d models? Thank you in advance!

— Reply to this email directly, view it on GitHub <

https://urldefense.com/v3/__https://github.com/Vision-CAIR/3DCoMPaT-v2/issues/11__;!!Nmw4Hv0!xQTKcPVQlbj7rHFFbYfin5W23H6Fd5hCxTHILNVPrim-xNAjwveimVnhkIRPpCqGtwR0MfDQfuzae8FUx3ptm1Go8oih1g$>,

or unsubscribe <

https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AG5EQIODQE2P7ITUNY6OSLT2ANYHVAVCNFSM6AAAAABRWZ2PJSVHI2DSMVQWIX3LMV43ASLTON2WKOZSGY2TKOJVGM4TIMQ__;!!Nmw4Hv0!xQTKcPVQlbj7rHFFbYfin5W23H6Fd5hCxTHILNVPrim-xNAjwveimVnhkIRPpCqGtwR0MfDQfuzae8FUx3ptm1E1eVHsMQ$>

. You are receiving this because you are subscribed to this thread.Message ID: @.***>

--

This message and its contents, including attachments are intended solely for the original recipient. If you are not the intended recipient or have received this message in error, please notify me immediately and delete this message from your computer system. Any unauthorized use or distribution is prohibited. Please consider the environment before printing this email.

— Reply to this email directly, view it on GitHub < https://github.com/Vision-CAIR/3DCoMPaT-v2/issues/11#issuecomment-2474013437 https://urldefense.com/v3/__https://github.com/Vision-CAIR/3DCoMPaT-v2/issues/11*issuecomment-2474013437__;Iw!!Nmw4Hv0!2OF-Ai1vHy_mPjS1zJvTlLPC5wRpRdPz1Q4zPivlo-moxjsJupH1xOo10V9uvN5RrLw0OA3Nqh6tsSYGfOWabcbhty8czw$>,

or unsubscribe < https://github.com/notifications/unsubscribe-auth/AW4ZEWB6FAGOYHJUILZFV432ANYURAVCNFSM6AAAAABRWZ2PJSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINZUGAYTGNBTG4 https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AW4ZEWB6FAGOYHJUILZFV432ANYURAVCNFSM6AAAAABRWZ2PJSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINZUGAYTGNBTG4__;!!Nmw4Hv0!2OF-Ai1vHy_mPjS1zJvTlLPC5wRpRdPz1Q4zPivlo-moxjsJupH1xOo10V9uvN5RrLw0OA3Nqh6tsSYGfOWabcbeqyTuEw$>

. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://urldefense.com/v3/__https://github.com/Vision-CAIR/3DCoMPaT-v2/issues/11*issuecomment-2474045898__;Iw!!Nmw4Hv0!2OF-Ai1vHy_mPjS1zJvTlLPC5wRpRdPz1Q4zPivlo-moxjsJupH1xOo10V9uvN5RrLw0OA3Nqh6tsSYGfOWabcbJR8ly_Q$, or unsubscribe https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AG5EQIJMKTADA4XV52SQLYD2AN2DJAVCNFSM6AAAAABRWZ2PJSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINZUGA2DKOBZHA__;!!Nmw4Hv0!2OF-Ai1vHy_mPjS1zJvTlLPC5wRpRdPz1Q4zPivlo-moxjsJupH1xOo10V9uvN5RrLw0OA3Nqh6tsSYGfOWabcYjsCMjjw$ . You are receiving this because you commented.Message ID: @.***>

--

This message and its contents, including attachments are intended solely for the original recipient. If you are not the intended recipient or have received this message in error, please notify me immediately and delete this message from your computer system. Any unauthorized use or distribution is prohibited. Please consider the environment before printing this email.