Saved in:
Bibliographic Details
Main Authors: Alawadh, Mona, Niranjan, Mahesan, Kim, Hansung
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.01431
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909412047716352
author Alawadh, Mona
Niranjan, Mahesan
Kim, Hansung
author_facet Alawadh, Mona
Niranjan, Mahesan
Kim, Hansung
contents Semantic Scene Completion (SSC) is a critical task in computer vision, that utilized in applications such as virtual reality (VR). SSC aims to construct detailed 3D models from partial views by transforming a single 2D image into a 3D representation, assigning each voxel a semantic label. The main challenge lies in completing 3D volumes with limited information, compounded by data imbalance, inter-class ambiguity, and intra-class diversity in indoor scenes. To address this, we propose the Multi-Feature Data Balancing Network (MDBNet), a dual-head model for RGB and depth data (F-TSDF) inputs. Our hybrid encoder-decoder architecture with identity transformation in a pre-activation residual module (ITRM) effectively manages diverse signals within F-TSDF. We evaluate RGB feature fusion strategies and use a combined loss function cross entropy for 2D RGB features and weighted cross-entropy for 3D SSC predictions. MDBNet results surpass comparable state-of-the-art (SOTA) methods on NYU datasets, demonstrating the effectiveness of our approach.
format Preprint
id arxiv_https___arxiv_org_abs_2412_01431
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Semantic Scene Completion with Multi-Feature Data Balancing Network
Alawadh, Mona
Niranjan, Mahesan
Kim, Hansung
Computer Vision and Pattern Recognition
Semantic Scene Completion (SSC) is a critical task in computer vision, that utilized in applications such as virtual reality (VR). SSC aims to construct detailed 3D models from partial views by transforming a single 2D image into a 3D representation, assigning each voxel a semantic label. The main challenge lies in completing 3D volumes with limited information, compounded by data imbalance, inter-class ambiguity, and intra-class diversity in indoor scenes. To address this, we propose the Multi-Feature Data Balancing Network (MDBNet), a dual-head model for RGB and depth data (F-TSDF) inputs. Our hybrid encoder-decoder architecture with identity transformation in a pre-activation residual module (ITRM) effectively manages diverse signals within F-TSDF. We evaluate RGB feature fusion strategies and use a combined loss function cross entropy for 2D RGB features and weighted cross-entropy for 3D SSC predictions. MDBNet results surpass comparable state-of-the-art (SOTA) methods on NYU datasets, demonstrating the effectiveness of our approach.
title Semantic Scene Completion with Multi-Feature Data Balancing Network
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.01431