Export a complete dataset to VSRS format, generating YUV images, and YUV disparity maps.
vsrs/export_for_vsrs.py dataset_parameters.json [simulate] [dataset_group]
dataset_parameters.json is a dataset parameters file for a dataset. If simulate is set, nothing is written, and only a simulation run is made. dataset_group can be set to a group of the dataset. Otherwise the default group is used.
Inside the script, variables can be set to indicate whether images and/or depth should be generated (image, depth), and if they should be overwritten if they already exist (overwrite_image, overwrite_depth).
The dataset parameters must contain image_filename_format and depth_filename_format for the vsrs group, and for the input (default or dataset_group) group. If must also contain z_far and z_near in the vsrs group.
The program will for each view index in the dataset (1D or 2D), read the input image and depth files, and write YUV image and disparity files into the path from the vsrs group.
If the option 16_bit_disparity in the vsrs group is set to true, then 16 bit disparity maps are generated. If not (by default) they are 8 bit.
Depth maps are converted to YUV disparity maps using vsrs/vsrs_disparity, with the z_far and z_near values from the dataset parameters. Images are converted from the original files to YUV420 files using tools/misc/yuv_export. The batch process can be parallelized.