Image annotations not being recognized onupload

Community Help
1

Project Type: Image Segmentation
OS & Browser: Ubuntu 22, Chrome
Project: Roof Obstructions Object Detection Dataset by Akhil

I’m trying to upload some png images, and a coco annotation file to an existing project with some jpeg images.

The images upload fine, but the annotations do not. The annotations file (annotations.json) looks like the following:

{
    "images": [
        {
            "id": 0,
            "width": 1002,
            "height": 1001,
            "file_name": "mFZD8h.png"
        },
        ...
    ],
    "annotations: [
        {
            "id": 0,
            "image_id": 0,
            "category_id": 0,
            "bbox": [
                339,
                555,
                14,
                12
            ],
            "area": 168,
            "segmentation": [
                [
                    342.7118449203307,
                    555.3099906005905
                ],
                [
                    339.9534230258326,
                    562.8026075213653
                ],
                [
                    350.84645285235854,
                    566.8128982233492
                ],
                [
                    353.60487474685664,
                    559.3202813025745
                ]
            ]
        },
        ...
    ],
    "categories": [
        {
            "id": 0,
            "name": "category_1"
        },
        ...
    ]
}

There are no licences in this json, because I generated it myself. Is there something wrong with my annotations, or have I done something else wrong?

2

Based on the information provided, it seems like you’re interested in starting an image segmentation project on Roboflow. There are two types of segmentation projects you can create: Instance Segmentation and Semantic Segmentation.

  • Instance Segmentation: This project type allows you to find the location of objects in an image at the pixel level. Each object found is given a unique reference.
  • Semantic Segmentation: This project type also allows you to find the location of objects in an image at the pixel level, but it does not differentiate between instances of the same object.

To create a project:

  1. Go to the Roboflow dashboard.
  2. Click “Create New Project”.
  3. Specify the project type (Instance Segmentation or Semantic Segmentation), the project name, and what you are detecting.

Once you’ve created your project, you can upload images via the web interface. If you already have a project, click “Upload” in the project sidebar to upload images. You can also clone images from Roboflow Universe to your project using the “Clone Images” feature.

For more detailed instructions, refer to the Roboflow documentation.

Please note that the operating system you’re using doesn’t affect the process of creating a project or uploading images on Roboflow, as it’s a web-based platform.

Hi! I’m Roboflow’s AI powered bot. This is an automated response powered using AI with knowledge from our docs, blog, and forum. If this solves your issue, please mark it as the solution using the checkbox icon. If this doesn’t solve your issue, please follow up with what you’ve tried/why, and the community will continue to chime in as usual.

3

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.