Image Processing & Analysis
Examples of image processing results produced at the BIOP. Top left: Wound Healing, Vignetting (Flatfield) correction, Deconvolution. Bottom left: Machine learning segmentation, deep learning segmentation with STARDIST, Simple segmentation with ImageJ.
As soon as you have finished the training on any microscope, or on request, you are granted access to our Image Processing and Analysis PCs in the main BIOP Office, AI-0140. You can access this room at any moment of the week with your CAMIPRO card.
Hardware
We use a variety of tools, open source and commercial , to help address you image analysis needs. Please get in touch with us with your biological question for a discussion or software training.
Workstations with 10Gb Fiber
We offer access to powerful workstations running on Windows 10 that can be booked using SV-BBS. These workstations are connected to EPFL servers with optic fiber, allowing transfer speeds of up to 500 MB/s.
Find the list of available workstations below, with their respective specs. You will also find the Operator 
and No Operator 
rates.
| PC Name | RAM | Processor | Graphics Processor | Notes | CHF | CHF |
|---|---|---|---|---|---|---|
| Peacock Spider | 128 GB | 14-core i9 3.1 GHz | 11 GB RTX 2080 Ti | 6.00 | 3.00 | |
| Bird of Paradise | 512 GB | 32-core AMD 5975WX 3.6 GHz | 24 GB RTX 3090 | Wacom Tablet | 6.00 | 3.00 |
| Orcinus Orca | 512 GB | 24-core AMD 5965WX 3.8 GHz | 24 GB RTX 3090 Ti | 6.00 | 3.00 | |
| Fluorescent Platypus | 256 GB | 10-core XEON 3.70 Ghz | 24 GB Titan RTX | Huygens Pro Licence | 6.00 | 3.00 |
Deconvolution Server
We offer a free of charge deconvolution service using SVI Huygens hosted on our PTBIOPSRV1 Linux Server. Access it through https://hrm-biop.epfl.ch
Server Specifications: 384 GB RAM, 48 AMD x64 cores @ 2.3 GHz
Software
We work with the most common microscopy and bioimage analysis software, both open source and commercial. We provide training sessions for most of these tools or train people on a per-project basis. Please do not hesitate to get in toutch with us.
Open Source Software
Commercial Software
Deep Learning
Since 2018, the BIOP image analysts have begun their foray and adoption of the many excellent and easy to use deep learning tools developed by groups around the world.
We can provide support and offer advice in the following software tools and packages:
Deep Learning Tools
Code, Community & Open Source
While we offer commercial image analysis tools to our users, most of our output is open source using Fiji/ImageJ and friends software jungle.
GitHub
Projects not linked to particular users or published works are made available through our GitHub account at https://github.com/BIOP.
EPFL Wikis
Most of our documentation, internal projects and repositories are available on our BIOP Wiki Page.
Image Processing Rules
- Long calculations? Try overnight or on weekends.
- Always switch off the Wacom tablets and monitors off when done.
- If a disk is full, data will be deleted indiscriminately.
Remote Access to PCs
Once you are able to use our workstations, we can enable remote desktop connections for your EPFL account. This is done on a per-user basis and requires that you acknowledge and follow these rules:
- Only log in when you have booked the system through SV-PPMS.
- Always log out (sign out) when you are finished (Don’t use ‘disconnect’).
- If you notice someone is logged in while trying to connect, call the BIOP to fix the situation.