Command line tools

Here are some commandline tools which we have included in library to make the process easy for anyone working on image classification or object detection and needs os related processing help like renaming files in sequence or divide data by folders etc.

Tools

Split Data

Most useful with image classification , split data as per train val test and QA with specific ratio.
note: if you just want to split data for train and test , set ratio 0 for that folder

split_data -d path_to_data_folder -r ratio_in_tuple_string
split_data -d "./cats_vs_dogs" -r "(70,10,10,10)"

This will split data inside class folder as 70% for training , 10% validiation, 10% testing and 10% for QA.

For more information about using command line tool:

split_data -h


Rename multiple files

Sequentially rename multiple files at once. You can also specify your own numbering series while renaming files

rename_files -n common name -f folder path -i number from which renaming to be started
rename_files -n "image_" -f "./image_folder" -i 1

This will rename the files in the image_folder

For more information about using command line tool:

rename_files -h


Tumor data extractor

Specifically for extracting and converting the mat files of brain tumor image data from the downloaded zip file using this link

tumor_data_extractor -b folder path of the downloaded zip folder -e extension in which mat files to be converted (default - jpg)
tumor_data_extractor -b "./1512427.zip" -e png

This will convert all the .mat file in the above zip folder into the given extension file format(default .jpg) All the converted files will be save in separate folder named brain_tumor_data in their respective tumor name folder

For more information about using command line tool:

tumor_data_extractor -h


Split data (GUI)

gui version of split data

split_data_gui