Ruman Ge s 1,2, Zol án Cse esnyés1, and Ma c Thilo Figge1,3
1 Applied Sys ems Biology, Leibniz Ins i u e o Na u al P oduc Resea ch and In ec ion Biology – Hans-Knöll-Ins i u e, Jena, Ge many
2 Facul y o Biological Sciences, F ied ich-Schille -Uni e si y Jena, Ge many
3 Ins i u e o Mic obiology, Facul y o Biological Sciences, F ied ich-Schille -Uni e si y Jena, Ge many
🗹 CLIJ2 algo i hms o GPU-accele a ed image p ocessing
🗹 Deep lea ning ia Cellpose
🗹 Soon: Omnipose suppo
🗹 Soon: Tenso flow in eg a ion
Re e ences
JIPipe: Designing au oma ed image analysis
pipelines wi hou p og amming
uman.ge s @leibniz-hki.de
Ja a Image P ocessing Pipeline (JIPipe, h ps://www.jipipe.o g/) [1]
is a plugin o ImageJ [2] ha allows o c ea e image p ocessing
pipelines wi hou a sc ip language. Ins ead use s only ha e o
design a flow-cha by connec ing p ocessing s eps (nodes) o each
o he in a isual p og amming language.
🗹 1000+ commonly used unc ions om ImageJ and popula plugins
🗹 In ui i e and mode n use in e ace
🗹 Powe ul anno a ion sys em o acking me ada a
🗹 S anda dized and au oma ed ou pu o ha d d i e
JIPipe in eg a es plen y o image analysis unc ions om ImageJ,
such as image impo , h esholding, con as enhancemen , edge
de ec ion, o ex ac ing measu emen s. Addi ionally, we p o ided
suppo o popula plugins including ...
To allow e en g ea e flexibili y, use s can u ilize ImageJ mac o,
Py hon, and R sc ip nodes o w i e o e-use cus om code inside
he JIPipe en i onmen .
Image analysis powe ed by ImageJ
GPU p ocessing and deep lea ning
Addi ional ea u es ia plugins
JIPipe allows easy u iliza ion o s a e-o - he-a image analysis
algo i hms ia i s au oma ed da a managemen capabili ies. The
cu en e sion includes he unc ions p o ided by he powe ul
CLIJ2 [3] lib a y, as well as he powe ul deep lea ning pla o m
Cellpose [4].
JIPipe al eady p o ides many ea u es, including nodes o able
p ocessing, plo ing, so ing and dis ibu ing da a, use in e ac ion
du ing he pipeline, and mo e.
Addi ional ea u es can be easily de eloped in Ja a as ImageJ
plugin, o c ea ed wi hin a GUI ool ha allows anyone o c ea e
new nodes. These a e also au oma ically a ailable om inside
ImageJ, as JIPipe expo s all i s unc ions o ImageJ.
JIPipe comes wi h i s own g aphical use
in e ace ha allows easy access o all
unc ions. ① The g aph edi o o build he flow
cha ② All se ings a e shown on he igh -hand
side
① ②
JIPipe al eady in eg a es many popula so wa e
ools like ImageJ and Py hon in o one easy- o-
use en i onmen .
JIPipe comes p e-packaged wi h suppo o GPU
and Deep Lea ning ools like CLIJ and Cellpose.
🗹 Bio-Fo ma s
🗹OMERO
🗹 Mo phoLibJ
🗹 Fea u eJ
🗹 Mul i-Templa e-Ma ching
🗹 T ainable Weka segmen a ion
🗹 Soon: T ackma e
🗹 Soon: Coloc2
Plugins
GUI Edi o
Ja a API
JIPipe can be easily ex ended wi h plugins.
🗹 Nodes
🗹 Da a ypes
🗹 Menus
🗹 ...
[1] Ge s e al. 2022. Resea ch Squa e. PREPRINT s.3. s-1641739/ 1
[2] Rueden e al. 2017. BMC Bioin o ma ics. 18(1): 1-26.
[3] Haase e al. 2020. Na Me hods. 17(1): 5-6.
[4] S inge e al. 2021. Na Me hods. 18(1): 100-106.
www.leibniz-hki.de
