DeePathology® STUDIO
Rapid Al creation made available for every pathologist
Powerful Al algorithms
The DeePathology@ STUDl0 is an interactive Do it Yourself platform for Al in pathology. Researchers and pathologists cancreate powerful Al solutions for histological image
analysis and use these algorithms to create powerful analytics on newpathology slides, Create proof of concepts (PoC) for most problems - in 5 minutes, no computer coding
required!
State-of-the-art neural networks
We are leveraging state of the art neural networks and deep leorning algorithms to make sure that you can create the mostpowerful solutions in the shortest time with the
lowest omount of doto
Lightning-fast annotation
Do you want to annotate thousands of cells in just a few days? With the DeePathology@ STUDl0, you can achieve this goalusing our innovative "Deep Interactive Online
Active Learning”.
High data security
The DeePathology@ STuDl0 is easy to use and fun while keeping your data privote. lt is installed on your local computer orlaptop, so you retain full control over your
sensitive data without having to send it to a remote server.
Versatile and flexible
Detect and classify cells and other objects, segment tissue regions and classify image tiles - all in one product. Combinemodes to create multi-layered solutions, for example
precise cell quantifcation within specifc tissue regions.

DeePathology® STUDl0 Objects Mode
Object segmentation
Create algorithms that detect the boundaries of objectsand classify them into unlimited categories,
Optimized for small data
Annotate as few as four to six examples from eachcategory to get started. Then run the algorithm andcorrect mistakes -your proof of concept solution is readyin minutes.
Annotation magic
The DeePathology® STUDlO also finds objects for you anddisplays them in the gallery. The gallery leverages anextremely powerful Active Learning Engine, suggesting
youthe most important examples to annotate. All you have todo is to set their category and you will obtain a high-quality balanced dataset in no time.
DeePathology® STUDlO Cells Mode
Cell detection and classification
Create algorithms that detect and classify different typesof cells and nuclei, and then detect millions of them in newslides. The cells mode leverages our powerful built-in
CellDetection building blocks that let you get accuratedetections in challenging real-life scenarios.
Powerful metrics
Monitor the performance of your algorithm in real-timeyou automatically obtain basic validation analytics foreach epoch.
Annotation magic
Using the gallery, the DeePathology® STUDlO makes surethat you are annotating the most important cells. With theAl mode, you can annotate thousands of cells in just a
fewminutes. And it is not just the quantity of cells - these arehigh quality images that will make your dataset diverse,and improve your algorithms the most. All you have to
dois sit back, drink some coffee, and correct mistakes fromtime to time in the gallery.
DeePathology® STUDlO Regions mode
Region segmentation
Create region segmentation algorithms for detectingregions of different types, for example different tissues ortumor regions vs. healthy tissue.
Multilayered analyses
You can then detect these region types in new slides, andcombine the algorithm with other DeePathology® STUDlOmodes to count different types of cells and objects
withinthe different regions.
DeePathology® STUDlO Tiles Mode
Tile classification
Quickly annotate tiles (small squares) with single clicks inthe viewer.
Optimized for speed
Use the gallery to annotate thousands of tiles fromdifferent slides in just minutes. Annotating tiles is simple,fast, and lets you create powerful solutions in record time.
Highly Configurable
To adapt to challenging reallife scenarios, there are many more diferent settings and modes in DeePathology@ STUDl0, for changingmagnification, controlling tissue
detection, algorithms sensitivities and more.
APl for Super Users
The APl for the DeePathology@ STuDl0 lets you run the algorithms created with DeePathology@ STuDl0 from code.This allows connecting the solutions developed with
DeePathology@ STUDl0 to any analytic pipeline, and use DeePathology@STUDlO aenerated algorithms as part of a larger process, for examole to predict patient outcomes
This also lets you run DeePathology@ STUDl0 generated algorithms on huge scales on your HPc to quickly deploy these algorithmson thousands of slides
One Al Platform -Countless Pathology Use Cases
H&E
Quantify tumor infiltrating lymphocytes (TlLs)
Segment tissue in tumor and healthy tissue
Detect glandular structures, such as glomeruli, acini, ..
Quantify mitotic figures
Segment the growth plate region in bone sections
Detect red blood cells
Seoment luno tissue and quantify cells in wall regions
Detect granulated and degranulated cells in skin cancer
Detect regions with high lymphocyte infiltration intumor sections
IHC
Quantify Ki67 positive cells and calculate the stain ratio
Segment positive and negative regions, for example in FoxP3, CD7and other immunostains
Detect different types of beta-amyloid plaques in Alzheimer'sdisease reserch
Quantify glia cells in neuroscience research, including astrocytesand microglia
Other stains
Quantify cells of interest in in-situ hybridization (lSH) stains
Detect Helicobacter pyloriin H&E or Giemsa stained slides
Detect regions infected with Mycobacterium tuberculosis in Ziehl-Neelsen stoined slides for veterinary TB surveillance

非常抱歉,您只有购买软件后才能查看完整软件教程!