convert_caltech_annos_to_yolo's Introduction

convert_caltech_annos_to_yolo

This script will convert caltech annotations into yolo format

You first need to extract .seq files into images and annotations into text files using either

evaluation tool from http://www.vision.caltech.edu/Image_Datasets/CaltechPedestrians/ and extract seq into images directory using dbExtract.m
with python https://github.com/mitmul/caltech-pedestrian-dataset-converter

next use convert_caltech_annos_to_yolo script to convert annotations into yolo format by providing

parser.add_argument('-images_root', default = 'F:\dataset\CaltechPedestrians\\train_images', # This dir name I cannot change cuz it will cause confusion 
                        help='path to root of the images dir\n', )
    parser.add_argument('-annos_root', default = 'F:\dataset\CaltechPedestrians\\txt_annotations',  
                        help='path to root of the annotations dir ')
    parser.add_argument('-yolo_annos_root', default = 'F:\\dataset\\CaltechPedestrians\\fullBB-sample\\train_labels', 
                        help='path to root of simple annotations dir')
    parser.add_argument('-filelist', default = 'F:\\dataset\\CaltechPedestrians\\fullBB-sample\\filelist_train.txt', 
                        help='name of the file that contain paths to images.')

convert_caltech_annos_to_yolo's People

Contributors

Stargazers

Watchers

convert_caltech_annos_to_yolo's Issues

Does not consider all images, only few are selected

Hi, thanks for the script.

I have 4250 images from the training dataset with their corresponding annotation files. I used Matlab toolkit to extract the .jpg and .txt files.
When I run the script it outputs the annotations for few files only. I changed the variable "skip" to "1", still annotations for only 1592 images are generated.
How to generate the annotations for every image?
Please help. Thanks a lot in advance.

Akshay

Recommend Projects