When annotating entities, I want to
. This makes it easier for me to compare the image with the text for entity annotation. How can I set this up to show the text along with its corresponding image?
. This makes it easier for me to compare the image with the text for entity annotation. How can I set this up to show the text along with its corresponding image?
This will require a custom recipe that uses the blocks interface. In particular you'll want to combine the ner_manual interface with the image interface.
Out of curiosity, why is the image required to annotate the text? Do you really need it? I'm mainly asking because it's usually a good idea to keep the annotation interfaces as simple as possible and I'm curious to hear about a use-case where adding the image provides helpful context.
- "It's much easier for me to annotate text when I have an image to refer to. If I only have the text, I can only annotate it section by section. (This is relatively simple to implement.)
- Currently, there are more convenient tools available. For instance, when I annotate directly on an image, the corresponding section of the text gets annotated automatically, and vice versa. This is more complex to implement. While the first method can still be modified and implemented in tools like Prodigy, the second method isn't supported yet and still requires significant manual effort. Other similar products include: UBIAI Easy to Use Text Annotation Tool | Create NLP Model"
Thanks for sharing 