Alp’s annotation converter plugin “Detectnet/KITTI to YOLO”

Detectnet KITTI Yolo annotation label convert

Hello, welcome to just another annotation converter. It works under Fiji, reads Detectnet/KITTI annotation text files from a folder and saves the Yolo versions to a different folder.

Installation

Download: Alps YoloConv (14mar2017, Windows)

Download and open the archive file above. There will be a “plugins” folder in it. Just drag and drop this folder onto “Fiji.app” folder.

Run Fiji. You will see the converter plugin’s name (Alps YoloConv), towards the end of “Plugins” main menu. Done.

Usage

Your detectnet label files and corresponding image files must be in the same folder. Then another (preferably empty) folder is necessary. Thats all you need. When you run the plugin, It asks for Detectnet labels’ folder (Input files folder). Then it asks where to save Yolo formatted label files.

Just wait until it completes converting the whole directory. During conversion, you can see the name of each processed label file (“.txt”) in the ‘Log’ window of Fiji.

I tried my best to prevent possible problems due to file structures, folder structures etc, but, its best if you can put only the necessary jpeg and text files into input folder. Please dont hesitate to report any problems, below.

After completion, select all newly created label files and paste/overwrite them into input folder. Your folder is ready to be trained under Yolo !

Current version is tested under Windows 10 (x64) Fiji environment. The software is provided “as is”. Its your responsibility to take necessary measures to prevent data loss.

I hope you find this macro plugin useful,

Alper

 

Versions

11 thoughts on “Alp’s annotation converter plugin “Detectnet/KITTI to YOLO”

  1. Hi I get an error like the one below. Where am I making the mistake.

    Invalid argument in line 54.
    textlist=Array. slice(textlist,c,textlist);

    Like

  2. Hi,How can this error remove Index(1) error out of 0-0 range in line 93
    if (isNaN(parseFloat(columns[0])) && (parseFloat(columns[1]) >= 0)){

    Like

    1. Hi, please check your data first. Try conversion with a small group of images, a few maybe. Checking label files for extra enters or spaces etc. may also help. Its working well in my tests.

      Like

  3. Hello,

    I am trying to convert labels obtained by ALT into yolo format by this plugin but getting this error.

    Current file is: Backward_Fall_A (2) 001.txt
    There is a problem with the annotation file below.
    Please double-click to see & fix the prob. & save it:

    please tell me how i can fix this. thank you

    Like

    1. This means there’s a problem with the way the data presented in the file. Please double check if it fits the format. Also, though not sure about this, but please try the filenames without paranthesis. Its always best to make some tests with a few files before running the tool for the whole directory.

      Like

  4. There is at least one .txt file do not comly with Detectnetnet/KITTI format.

    I am receiving this message. my images in jpeg format and labels in .txt are in same folder which i am using as an input. I am unable to resolve this error message. Kindly guide

    Like

    1. Hi Ali, as I advised, try to run the tool with, say, 2 or 3 images and their label text files to see if tool is properly working. So arrange a folder with these files and run the tool. If it fails, try with only one image and its label text file. For me, its almost certain that a label text file is problematic. You need to spot it. Ensure you properly prepared the labels. Use “Labels to Images Converter” to check validity of your labels.
      Best,

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s