Instructions

The instructions for downloading CODA is listed as follows:

  1. Download the CODA dataset files using the link provided below and then decompress.
  2. Check the CODA Documentation page and the included README file for specific usages of CODA dataset, especially for the related data of KITTI and nuScenes datasets.

After decompression, the data organization is listed as follows: 

"CODA": {
    |---base-val-1500                           
        |---corner_case.json            -- COCO-style annotation
        |---kitti_indices.json          -- KITTI sample indices
        |---nuscenes_sample_tokens.json -- nuScenes sample tokens
        |---images                      -- ONCE images (also included in CODA2022 validation)
            |---*_*.jpg
        |---README.md                   -- Detailed instructions about CODA Base Val usage                 
    |---sample                          -- Proper subset of CODA base validation set 
        |---corner_case.json            
        |---images        
            |---*_*.jpg     
    |---val                             -- CODA2022 validation set 
        |---annotations.json            
        |---images        
            |---*_*.jpg     
    |---test                            -- CODA2022 test set 
        |---annotations.json            
        |---images        
            |---*_*.jpg     
    |---README.md                       -- Detailed instructions about CODA usage
}

Download Links

The ground truth annotation of CODA2022 test set will be released after the SSLAD challenge, and extraction code for Baidu Yunpan is 4wbz.

Download link for CODA sample set (~100 images): [Link]
Download link for CODA base val set: [Google Drive] [Baidu Yunpan]
Download link for CODA2022 val set (images with annotations): [Google Drive]
Download link for CODA2022 test set (images with annotations): [Google Drive]