The Scene type classification task involves classifying a scan into 13 scene types.

Evaluation and metrics

Our evaluation ranks all methods according to recall (TP/(TP+FN)) as well as the PASCAL VOC intersection-over-union metric (IoU = TP/(TP+FP+FN)), where TP, FP, and FN are the numbers of true positive, false positive, and false negative predictions, respectively.



This table lists the benchmark results for the scene type classification scenario.




Method Infoavg iouapartmentbathroombedroom / hotelbookstore / libraryconference roomcopy/mail roomhallwaykitchenlaundry roomliving room / loungemiscofficestorage / basement / garage
sorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
multi-taskpermissive0.646 10.500 11.000 10.789 10.333 20.667 21.000 10.500 11.000 11.000 10.778 10.000 20.833 10.000 2
Shengyu Huang, Mikhail Usvyatsov, Konrad Schindler: Indoor Scene Recognition in 3D. IROS 2020
3DASPP-SCE0.556 20.500 10.938 20.778 20.667 11.000 10.250 20.500 10.750 20.333 20.500 30.000 20.812 20.200 1
SE-ResNeXt-SSMA0.355 30.000 40.684 30.696 30.200 40.500 30.200 30.500 10.429 30.200 30.545 20.111 10.556 30.000 2
Abhinav Valada, Rohit Mohan, Wolfram Burgard: Self-Supervised Model Adaptation for Multimodal Semantic Segmentation. arXiv
resnet50_scannet0.231 40.200 30.481 40.346 40.250 30.250 40.000 40.500 10.333 40.000 40.357 40.000 20.286 40.000 2