I am now trying to use the prediction results from Mask3D in 3D visual grounding, same as what 3D-VisTA does.
However, in the github of 3D-VisTA, they only provide the predictions results of Mask3D in parts of the scenes of ScanNet v2. Mainly because the 3D visual grounding dataset, ScanRefer, just use parts of the scenes in ScanNet v2.
I am not familiar with this 3D segmentation field, and I know there's some preprocess and postprocess in your coding. Would you please do me a favor to directly provide the segmentation results of ScanNet? That would be a great help to me!
Dear authors, thanks for your great job!
I am now trying to use the prediction results from Mask3D in 3D visual grounding, same as what 3D-VisTA does.
However, in the github of 3D-VisTA, they only provide the predictions results of Mask3D in parts of the scenes of ScanNet v2. Mainly because the 3D visual grounding dataset, ScanRefer, just use parts of the scenes in ScanNet v2.
I am not familiar with this 3D segmentation field, and I know there's some preprocess and postprocess in your coding. Would you please do me a favor to directly provide the segmentation results of ScanNet? That would be a great help to me!