search

matterport / Mask_RCNN

Mask R-CNN for object detection and instance segmentation on Keras and TensorFlow
Other
24.53k stars 11.68k forks source link

Using the weights of a custom model in combination with Mask-RCNN #2793

Open aizzaac opened 2 years ago

aizzaac commented 2 years ago

Hi

I have the weights of a model which uses semantic segmentation in floor plan images to detect things like: walls, windows, type of rooms.

I have a new dataset of floor plan images and I want to train a model, using Mask-RCNN and the weights of the previous mentioned model.

Is that possible? Or do the annotations have to be equal (same number of instances)?

Thank you

Some details about the previous mentioned model: semantic segmentation classes for walls windows, types of rooms (kitchen,bathroom)

New dataset: annotation for each instance classes for walls, windows, doors, columns, types of rooms(kitchen, bedroom, bathroom, office, etc), office seat, sink, toilet, bike parking, etc.

caiqueaugustocm commented 2 years ago

Did you manage to solve it? I'm in similar problem, I have an .h5 file (custom weighs) and I can't use in custom dataset.