Show HN: Detecto – Computer vision model in 10 lines of code or less

by alanbi • 7 years ago

7 3 7 years ago

alanbi2• 7 years ago

Hey all!

I'm a university student and have been working on this project for the past month or two. I wanted to learn how to create computer vision/object detection models for my school's robotics club but found the learning curve pretty steep. With Detecto, my hope is that it makes the field more approachable to newcomers and gives people a quick and easy option for when the existing libraries are too heavy-duty.

Please let me know what you think and if you have any feedback!

pretty_dumm_guy• 7 years ago

Hi,

Awesome work. Something similar to what I did, though for semantic segmentation.

A small suggestion: From what I understand, you have hard-coded the Fast-RCNN predictor.

  self._model.roi_heads.box_predictor = FastRCNNPredictor(in_features, len(classes) + 1)

Having the ability to plugin users' own models would be better, I believe. This would translate to making the model class more abstract.

Well done!

alanbi2• 7 years ago

Thanks for the suggestion! That's a feature I definitely plan on adding sometime in the near future.

Show HN: Detecto – Computer vision model in 10 lines of code or less

Related Stories