SqueezeBrains SDK 1.18
|
Image tiling parameters. More...
#include <sb.h>
Data Fields | |
sb_t_tiling_mode | mode |
Tiling mode to use for image processing. More... | |
int | num_tiles |
Number of horizontal and vertical tiles used to process the image. More... | |
sb_t_range_flt | scale |
Image scale. More... | |
Image tiling parameters.
Tiling parameters used to dermine the policy with which images will be tiled (divided) by sb_svl_run and sb_project_detection functions. It is possible to set different tiling parameter configuration horizontaly and vertically.
sb_t_tiling_mode sb_t_svl_dl_tiling_par::mode |
Tiling mode to use for image processing.
"Tiling" refers to that image processing operation that consists in subdividing the elaborated image into multiple sub-images.
Used only by Deep Surface projects.
int sb_t_svl_dl_tiling_par::num_tiles |
Number of horizontal and vertical tiles used to process the image.
Used only by Deep Surface projects and if sb_t_svl_dl_tiling_par::mode is equal to SB_TILING_MODE_MANUAL.
Couple of values that determines a grid scheme used to subdivide, both horizontally and vertically, the original image into tile_factor.width * tile_factor.height tiled images. Each tile is processed by sb_svl_run and sb_project_detection function as a single image.
Applying a tile factor > {1, 1} is useful to increase the image resolution at the input of the elaboration algorithm, especially when the sb_t_svl_dl_par_network::input_size of the network is significantly lower than the resolution of the image. This may help to detect small defects instances and to have a more accurate segmentation at pixel levels.
On the other hand, the higher the number of tiles of the grid, the higher is the training and detection time and also the GPU memory usage of the detection.
Values range from SB_SVL_DL_NUM_TILES_MIN to SB_SVL_DL_NUM_TILES_MAX.
If it is necessary that the image is processed with a defined scale factor in order to guarantee a minimum defects, the tiling mode can be set to SB_TILING_MODE_AUTO in order to manually select the required scale.
Higher tile factor values on computational device with limited memory resources may cause SB_ERR_DL_CUDA_OUT_OF_MEMORY .
sb_t_range_flt sb_t_svl_dl_tiling_par::scale |
Image scale.
Range of scale to apply to the image when sb_t_svl_dl_tiling_par::mode parameter is equal to SB_TILING_MODE_AUTO. scale.max is used to compute the number of tiles used to divide the image. scale.min is used to determine how much increase the image resolution along a direction in order to maximaze the image area underlying of the tile.