Tool for annotation of images for semantic segmentation

Question

I have been searching around for a software tool, that I can use for annotating images. More specifically I want to do annotation to be used for semantic segmentation, meaning I want to create masks. I want to be able to create training data for applying a segmentation CNN (like for instance U-net).

However I have been digging around the internet, and I have tried out some options. But I have not really found anything that seems to do the trick.

I do not expect to have many classes, but I probably want to be able to be able to annotate more than two classes (something, background). And of course every pixel has to belong to a class.

So what I am looking for is a relatively simple tool that either lets you mark areas of a class with a brush or that lets one mark polygons and then fill those, and then the outside the polygons as background. So can anyone point me to anything like this? And share their experience?

Answer 1

Diffgram is really great for this! I used it for a construction monitoring project. It's Open Source. From their site:

Semantic Segmentation Tools:

• Auto Bordering: Automatically detects edges to create 100% coverage masks. Simple select the intersecting shape.
• Combo Shapes: Create shapes that are partially curves and partially straight lines.
• Points to Full Polygon: Draw a box (or a few points), get a polygon.
• Correct Pre-Label Segmentation Map: Load an existing segmentation prediction (pixels.) Get an editable polygon for correction.

Semantic Segmentation Data Annotation

Answer 2

You might look at: https://github.com/facebookresearch/segment-anything/tree/main

Iy works very well at segmenting all sorts of stuff out of the box. There are examples around showing how to fine-tune it for particular domains.