How to train a Faster R-CNN with a Multi-labeled images table?

9 views (last 30 days)
Claudio Eutizi
Claudio Eutizi on 14 Jan 2021
Edited: Jack Xiao on 22 Feb 2021
I have a 8792x11 table with my dataset labeled with the Image Labeler:
1st column: imageFilename;
from 2nd to 11th column: different labels with the rectangles' dimensions, exactly as how the ground-truth dataset must be for Faster R-CNN training.
But MathWorks' examples show how to train the network with a one-labeled dataset.
How to train the net with a multi-labeled ground-truth table?
Thank you.

Answers (1)

Jack Xiao
Jack Xiao on 22 Feb 2021
Edited: Jack Xiao on 22 Feb 2021
what does multi-labeled ground-truth mean? the MathWorks' rcnn example is for object detection with mutilple instances but one-labeld ground-truth.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by