Papers › Single Frame Semantic Segmentation Using Multi-Modal Spherical Images

Single Frame Semantic Segmentation Using Multi-Modal Spherical Images

18 Aug 2023arXiv:2308.09369archive 2025-07-28

Suresh Guttikonda, Jason Rambach

In recent years, the research community has shown a lot of interest to panoramic images that offer a 360-degree directional perspective. Multiple data modalities can be fed, and complimentary characteristics can be utilized for more robust and rich scene interpretation based on semantic segmentation, to fully realize the potential. Existing research, however, mostly concentrated on pinhole RGB-X semantic segmentation. In this study, we propose a transformer-based cross-modal fusion architecture to bridge the gap between multi-modal fusion and omnidirectional scene perception. We employ distortion-aware modules to address extreme object deformations and panorama distortions that result from equirectangular representation. Additionally, we conduct cross-modal interactions for feature rectification and information exchange before merging the features in order to communicate long-range contexts for bi-modal and tri-modal feature streams. In thorough tests using combinations of four different modality types in three indoor panoramic-view datasets, our technique achieved state-of-the-art mIoU performance: 60.60% on Stanford2D3DS (RGB-HHA), 71.97% Structured3D (RGB-D-N), and 35.92% Matterport3D (RGB-D). We plan to release all codes and trained models soon.

PaperPDFCode

Code

sguttikon/SFSS-MMSI officialmentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Semantic Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Semantic Segmentation Matterport3D SFSS-MMSI (RGB+Depth) Test mIoU 35.92 #1 of 4 Archive leaderboard report
Semantic Segmentation Matterport3D SFSS-MMSI (RGB+Depth) Validation mIoU 39.19 #1 of 4 Archive leaderboard report
Semantic Segmentation Matterport3D SFSS-MMSI (RGB+Normal) Test mIoU 35.77 #2 of 4 Archive leaderboard report
Semantic Segmentation Matterport3D SFSS-MMSI (RGB+Normal) Validation mIoU 38.91 #2 of 4 Archive leaderboard report
Semantic Segmentation Matterport3D SFSS-MMSI (RGB+Depth+Normal) Test mIoU 35.52 #3 of 4 Archive leaderboard report
Semantic Segmentation Matterport3D SFSS-MMSI (RGB+Depth+Normal) Validation mIoU 39.26 #3 of 4 Archive leaderboard report
Semantic Segmentation Matterport3D SFSS-MMSI (RGB Only) Test mIoU 31.3 #4 of 4 Archive leaderboard report
Semantic Segmentation Matterport3D SFSS-MMSI (RGB Only) Validation mIoU 35.15 #4 of 4 Archive leaderboard report
Semantic Segmentation Stanford2D3D Panoramic SFSS-MMSI (RGB+HHA) mAcc 70.68 #1 of 25 Archive leaderboard report
Semantic Segmentation Stanford2D3D Panoramic SFSS-MMSI (RGB+HHA) mIoU 60.6% #1 of 25 Archive leaderboard report
Semantic Segmentation Stanford2D3D Panoramic SFSS-MMSI (RGB+Depth+Normal) mAcc 69.03 #2 of 25 Archive leaderboard report
Semantic Segmentation Stanford2D3D Panoramic SFSS-MMSI (RGB+Depth+Normal) mIoU 59.43% #2 of 25 Archive leaderboard report
Semantic Segmentation Stanford2D3D Panoramic SFSS-MMSI (RGB+Normal) mAcc 68.79 #3 of 25 Archive leaderboard report
Semantic Segmentation Stanford2D3D Panoramic SFSS-MMSI (RGB+Normal) mIoU 58.24% #3 of 25 Archive leaderboard report
Semantic Segmentation Stanford2D3D Panoramic SFSS-MMSI (RGB+Depth) mAcc 68.57 #5 of 25 Archive leaderboard report
Semantic Segmentation Stanford2D3D Panoramic SFSS-MMSI (RGB+Depth) mIoU 55.49% #5 of 25 Archive leaderboard report
Semantic Segmentation Stanford2D3D Panoramic SFSS-MMSI (RGB Only) mAcc 63.96 #9 of 25 Archive leaderboard report
Semantic Segmentation Stanford2D3D Panoramic SFSS-MMSI (RGB Only) mIoU 52.87% #9 of 25 Archive leaderboard report
Semantic Segmentation Structured3D SFSS-MMSI (RGB+Depth+Normal) Test mIoU 71.97 #1 of 4 Archive leaderboard report
Semantic Segmentation Structured3D SFSS-MMSI (RGB+Depth+Normal) Validation mIoU 75.86 #1 of 4 Archive leaderboard report
Semantic Segmentation Structured3D SFSS-MMSI (RGB+Normal) Test mIoU 71 #2 of 4 Archive leaderboard report
Semantic Segmentation Structured3D SFSS-MMSI (RGB+Normal) Validation mIoU 74.38 #2 of 4 Archive leaderboard report
Semantic Segmentation Structured3D SFSS-MMSI (RGB+Depth) Test mIoU 70.17 #3 of 4 Archive leaderboard report
Semantic Segmentation Structured3D SFSS-MMSI (RGB+Depth) Validation mIoU 73.78 #3 of 4 Archive leaderboard report
Semantic Segmentation Structured3D SFSS-MMSI (RGB Only) Test mIoU 68.34 #4 of 4 Archive leaderboard report
Semantic Segmentation Structured3D SFSS-MMSI (RGB Only) Validation mIoU 71.94 #4 of 4 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections