This is the PyTorch based code for the paper "Bounding Box Priors for Cell Detection with Point Annotations" https://arxiv.org/abs/2211.06104.
The paper proposes a weakly semi-supervised machine learning algorithm to detect objects with point annotations.
!python train.py --train-file train.csv
--workers 16 --batch-size 2 --epoch 50 --lr 0.01 --beta 0.99 --eval-freq 50
--device cuda --amp --balance
--gt-bbox-loss l1
--data-path ../data/BCCD/ --results-dir ../results --config bccd_wl100
Please send an email at hariom85@gmail.com.