FBMS-59

Freiburg-Berkeley Motion Segmentation

ImagesVideosIntroduced 2010-01-01

The Freiburg-Berkeley Motion Segmentation Dataset (FBMS-59) is a dataset for motion segmentation, which extends the BMS-26 dataset with 33 additional video sequences. A total of 720 frames is annotated. FBMS-59 comes with a split into a training set and a test set. Typical challenges appear in both sets.

Source: https://lmb.informatik.uni-freiburg.de/resources/datasets/moseg.en.html Image Source: https://lmb.informatik.uni-freiburg.de/resources/datasets/moseg.en.html

Benchmarks

16k/S-Measure 16k/MAX E-MEASURE 16k/MAX F-MEASURE 16k/AVERAGE MAE 2D Classification/S-Measure 2D Classification/MAX E-MEASURE 2D Classification/MAX F-MEASURE 2D Classification/AVERAGE MAE 2D Object Detection/S-Measure 2D Object Detection/MAX E-MEASURE 2D Object Detection/MAX F-MEASURE 2D Object Detection/AVERAGE MAE 3D/S-Measure 3D/MAX E-MEASURE 3D/MAX F-MEASURE 3D/AVERAGE MAE Instance Segmentation/mIoU Object Detection/S-Measure Object Detection/MAX E-MEASURE Object Detection/MAX F-MEASURE Object Detection/AVERAGE MAE RGB Salient Object Detection/S-Measure RGB Salient Object Detection/MAX E-MEASURE RGB Salient Object Detection/MAX F-MEASURE RGB Salient Object Detection/AVERAGE MAE Unsupervised Object Segmentation/mIoU Video/mIoU Video/S-Measure Video/MAX E-MEASURE Video/MAX F-MEASURE Video/AVERAGE MAE Video Object Segmentation/mIoU Video Object Segmentation/S-Measure Video Object Segmentation/MAX E-MEASURE Video Object Segmentation/MAX F-MEASURE Video Object Segmentation/AVERAGE MAE