This is a instruction about Autosweep Dataset.
- 1_ Cube_Cylinder_11658: Contains all Images, masks and etc
- cls: FCIS data, classes level
- gray: Contour maps
- ImageSets: Training lists,Testing lists
- img: Image...
- inst: FCIS data, instances level
- instanceIds_5classes: Masks of MaskRCNN data, 5 classes
- instanceIds_6classes: Masks of MaskRCNN data, 6 classes
- line: Edge maps
- lineData: For modeling, Record line Data.
- lineImage: For modeling, visulization line.
- seg: Masks of FCIS
- tool: Useful Tools
- old_cube_files: Old cube file with deficiency
- 2_Cube_Cylinder_raw: Backup data for deeplab, faster-rcnn
- ...
- 3_Synthtic_Data_21467: Synthtic Data with 10 models
- ...
- 4_Sub_Geonet_data: Training and Testing Data for Sub Geonet
- ...
- 5_XiaoData: XiaoData with Autosweep pattern in 1_ Cube_Cylinder_11658.
- 6_AxisCurveData: Training and Testing data for axis curve experiments
- 7_ExperimentData:
- Exper_Cuboid: Training Data for Cuboid Modeling and Detection Experiment
- Exper_FCIS_DCN: Training Data for MeanAP Experiment
- Exper_MeanAP: Testing Data for MeanAP Experiment, For evaluation
- Exper_PointNet_VolexNet: models for PointNet & VolexNet
- Exper_XiaoCube: Testing Data for Cuboid Modeling and Detection Experiment, For evaluation
- Net_Model: Config files and Models for FCIS, MaskRcnn & DCN
- Rebuttal_Exper_Mask_RCN: Rebuttal Experiments on MeanIOU.