English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Cell detection is an important task in biomedical research. Recently, deeplearning methods have made it possible to improve the performance of celldetection. However, a detection network trained with training data under aspecific condition (source domain) may not work well on data under otherconditions (target domains), which is called the domain shift problem. Inparticular, cells are cultured under different conditions depending on thepurpose of the research. Characteristics, e.g., the shapes and density of thecells, change depending on the conditions, and such changes may cause domainshift problems. Here, we propose an unsupervised domain adaptation method forcell detection using a pseudo-cell-position heatmap, where the cell centroid isat the peak of a Gaussian distribution in the map and selectivepseudo-labeling. In the prediction result for the target domain, even if thepeak location is correct, the signal distribution around the peak often has anon-Gaussian shape. The pseudo-cell-position heatmap is thus re-generated usingthe peak positions in the predicted heatmap to have a clear Gaussian shape. Ourmethod selects confident pseudo-cell-position heatmaps based on uncertainty andcurriculum learning. We conducted numerous experiments showing that, comparedwith the existing methods, our method improved detection performance underdifferent conditions.

Effective Pseudo-Labeling based on Heatmap for Unsupervised Domain Adaptation in Cell Detection