Register now After registration you will be able to apply for this opportunity online.
3D Object Instance Segmentation
The project focuses on identifying distinct object instances (such as tables and chairs) within 3D reconstructions of indoor environments.
Keywords: 3D instance segmentation, reconstruction, semantics
The project focuses on segmenting distinct object instances (such as tables and chairs) in 3D reconstructions of indoor environments. It primarily aims to enhance networks that rely solely on 3D point data (e.g., Mask3D [a]) by integrating the textural and fine details available from images with the geometric information contained in the 3D point clouds. This hybrid approach is expected to improve the accuracy and reliability of object detection in complex spatial settings, focusing especially on small objects and parts of bigger objects.
[a] https://francisengelmann.github.io/mask3d/index.html
The project focuses on segmenting distinct object instances (such as tables and chairs) in 3D reconstructions of indoor environments. It primarily aims to enhance networks that rely solely on 3D point data (e.g., Mask3D [a]) by integrating the textural and fine details available from images with the geometric information contained in the 3D point clouds. This hybrid approach is expected to improve the accuracy and reliability of object detection in complex spatial settings, focusing especially on small objects and parts of bigger objects.