Back

A Weakly Supervised U-Net Model for Precise Whole Brain Immunolabeled Cell Detection

Wang, L.-W.; Wu, Y.-L.; Lee, C.-L.; Cheng, C.-C.; Lu, K.-Y.; Tsai, J.-H.; Lin, Y.-H.; Hsu, C.-H.; Kuo, T.-H.; Chu, L.-A.

2023-03-18 neuroscience
10.1101/2023.03.16.531434 bioRxiv
Show abstract

Cell segmentations low precision due to the intensity differences hinders widespread use of whole brain microscopy imaging. Previous studies used ResNet or CNN to account for this problem, but are unapplicable to immunolabeled signals across samples. Here we present a semiauto ground truth generation and weakly-supervised U-Net-based Deep-learning precise segmentation pipeline for whole brain immunopositive c-FOS signals, which reveals the distinct neural activity maps with different social motivations.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.