make-sense
make-sense copied to clipboard
SkalskiP
User tour & guide
The experience of the project is great. For improvements, I am thinking it will be good to have a tour prompt that displays next to the important actions, especially as it concerns the following:
- Adding Labels: A prompt next to the + button and disappears once user clicks on the button
- Project Title: A prompt on setting the project title, disappears once user click to edit
- Export: A prompt showing the user the export button after they are done annotations.
See the link below to an example prompt tool from bootstrap.
https://bootstraptour.com/
@OlafenwaMoses Thank you for another great idea. Such a trip could definitely help new users in particular. Surely I would like to have this feature in the application. However, this is something that will probably land a little further in the timeline. There are some features that I really care about. First of all labeling with polylines and keyboard shortcuts.
Just adding my 2 cents here, How about adding a practical example (even just in the readme) to explain what the benefit is of this app? I am not quite certain I understand the practical value of this app? Is there anything automatic this gives me besides a csv output?
@hitautodestruct I think that in the near future I will start preparing articles on Medium, showing how to use makesense.ai to prepare a dataset for model training. Parallel to the development of UI and the addition of various functionalities facilitating the work with the editor, some AI functionalities are being developed. They are supposed to support the work with the editor. I hope that they will get to the editor as soon as possible.
@hitautodestruct In total, we have a few ideas for automating the labeling process:
- We are thinking about placing ready-made, trained models (transferred to JS), which would of course be able to recognize only a set of classes selected in advance. I wonder if such a solution makes sense. It may be useful if our user has a data set with objects, some of which would be supported by us. Then we would help him locate some of them automatically, and he would have to add the others.
- We are also thinking about a solution based on transfer learning in JS, but it is quite a complicated idea. Because it would require a lot of work from us and we don't know when we will be able to implement it.
- There was also an idea to create a solution based on a traditional CV, which would locate some objects, not knowing exactly what kind of object it is, but allowing it to surround the bounding box.