• Compile code

  1. Download CUDA 7.5 and cuDNN 3. You will need to register with NVIDIA.
  2. cd code/marvin
  3. ./linux.sh

• Prepare data download the processed RGBD data : http://dss.cs.princeton.edu/Release/sunrgbd_dss_data
  by runing script: downloadData('../sunrgbd_dss_data','http://dss.cs.princeton.edu/Release/sunrgbd_dss_data/','.bin'); or run dss_preparedata() to prepare your own data.

  download the image and hha images: http://dss.cs.princeton.edu/Release/hha http://dss.cs.princeton.edu/Release/image by runing script: downloadData('../image','http://dss.cs.princeton.edu/Release/image/','.tensor');

• 3D region proposal network:

  1. You can download the precomputed region proposal for NYU and SUNRGBD dataset from: http://dss.cs.princeton.edu/Release/result/proposal/RPN_SUNRGBD/ http://dss.cs.princeton.edu/Release/result/proposal/RPN_NYU/ by runing script: downloadData('../proposal','http://dss.cs.princeton.edu/Release/result/proposal/RPN_NYU/','.mat');

  2. To train 3D region proposal network and extract 3D region proposal cd code/matlab_code/slidingAnchor run dss_prepareAnchorbox() to prepare training data. run RPN_extract() to extract 3D region proposal. You may need the segmentation result here: downloadData('../seg','http://dss.cs.princeton.edu/Release/seg/','.mat');

• 3D object detection network:

  1. change path in dss_initPath.m;
  2. run dss_marvin_script(0,100,1,[] ,1,'RPN_NYU',1,[],0,0); Pretrained model:
     http://dss.cs.princeton.edu/Release/pretrainedModels/DSS/ORN/DSSnet_ORN_d.marvin

Notes : 1. If matlab system call fails, you can try to run the command directly. 2. The rotation matrixes for some of the images in the dataset are different from the original SUNRGB-D dataset, so that the rotation only contains camera tilt angle (i.e. point cloud does not rotated on the x,y plane). We provide the data in this repo "./external/SUNRGBDtoolbox/Metadata/SUNRGBDMeta.mat" All the results and ground truth boxes provided in this repo are using this rotation matrix. To convert the rotation matrix you can reference the code "changeRoomR.m"