Back

EzReverse - a Web Application for Background Adjustment of Color Images

Song, X.; Goedhart, J.

2024-05-27 bioinformatics
10.1101/2024.05.27.594095 bioRxiv
Show abstract

Fluorescence imaging is extensively utilized in biological research to study biomolecules at the single-cell level. Fluorescence images have a black background due to the inherent nature of fluorescence imaging. The black background fits well with dark themes that are used on screens and presentations. On the other hand, inverting the color images to display the signals on a white background can aid the human visual system in capturing detailed information or subtle features. In addition, white backgrounds match better with printed media. However, current methods to invert or change the background of color images require multiple steps or do not offer flexibility in the adjustment which is important when dealing with complex images without raw data. To facilitate easy access to background inversion and flexible modifications, ezReverse was developed. Two methods are implemented, the first based on color space transformation followed by inverting the lightness, while the second method identifies grayscale values which are subsequently modified. The web app ezReverse is a versatile tool, that accommodates multiple color spaces, kernel filters, and gamma correction for optimal inversion of color images. It is hosted on the Shiny Python platform and can be conveniently accessed online: https://amsterdamstudygroup.shinyapps.io/ezreverse/ O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=186 SRC="FIGDIR/small/594095v1_ufig1.gif" ALT="Figure 1"> View larger version (50K): org.highwire.dtl.DTLVardef@7a0901org.highwire.dtl.DTLVardef@1e53776org.highwire.dtl.DTLVardef@19c6593org.highwire.dtl.DTLVardef@d0a722_HPS_FORMAT_FIGEXP M_FIG C_FIG

Matching journals

The top 7 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.