#super-resolution

Fixed Energy

Like filming the bustle of a city at night, life inside a cell can look like chaos. A static snapshot allows us to make sense of the surroundings, to spot subtle details, and perhaps some order in the noise. In these mammalian cells, we see clusters of mitochondria, the tiny organelles that churn out cellular energy, a molecule called ATP. Among other jobs, mitochondria also heat our cells from the inside – each is around a billion times smaller than a nuclear power facility. While watching organelles in motion has its advantages, here a new fluorescent stain reveals fine details in fixed cells – those frozen in time with chemicals – potentially yielding different perspectives on these tiny power stations in health and disease.

Written by John Ankers

Image from work by Jingting Chen and colleagues

College of Future Technology, Institute of Molecular Medicine, National Biomedical Imaging Center, Beijing Key Laboratory of Cardiometabolic Molecular Medicine, Peking University, Beijing, China

Image originally published with a Creative Commons Attribution 4.0 International (CC BY 4.0)

Published in Proceedings of the National Academy of Science (PNAS), April 2024

Despite remarkable developments in diffraction unlimited super-resolution microscopy, in vivo nanoscopy of tissues and model organisms is still not satisfactorily established and rarely realized. RESOLFT nanoscopy is particularly suited for live cell imaging because it requires relatively low light levels to overcome the diffraction barrier.

Above : Drosophila melanogaster ubiquitously expressing rsEGFP2-α-tubulin.

POLCAM Action!

POLCAM, a modification of a method for super-resolution microscopy called SMOLM by detecting polarisation with the addition of a polarising camera to any wide-field fluorescence microscope, plus open-source analysis software

Read the original research article here

Image from work by Ezra Bruggeman and colleagues

Yusuf Hamied Department of Chemistry, University of Cambridge, Cambridge, UK

Image originally published with a Creative Commons Attribution 4.0 International (CC BY 4.0)

Published in bioRxiv, May 2024 (not peer reviewed)

We strive to offer the best products & services available, such as high definition super-resolution image enlargement printing. As a forerunner in computer A.I. advanced algorithm use, our accumulated data & experience printing high resolution images to a large print format is unique. Visit our Reviews page to see what our customers think.

I figured out how to get ESRGAN (and SFTGAN) running on Windows. Unfortunately, it’s not as user friendly as waifu2x, but it’s fairly easy as far as Python deep learning programs go.

The most up to date info is at Reddit’s /r/gameupscale. https://www.reddit.com/r/GameUpscale/ You can also check out the Game Upscale Discord here: https://discord.gg/jkFHzGD 

I’m leaving this guide up because it still gets quite a bit of traffic.

You will need:

Windows 7 or up, 64-bit only

An Nvidia GPU is recommended. the newer and the more memory, the better. My 6GB 1060 could upscale 1024x600 images to 4096x2400. You’ll need to run it in CPU mode otherwise - see step 3.5.

Python 3 64-bit (x86-64) : https://www.python.org/downloads/release/python-366/

Step 1: Install CUDA 9.2 or 10.0, plus the patch if there is one. I believe you only need the CUDA and driver components: https://developer.nvidia.com/cuda-toolkit-archive

Step 2: Go to the PyTorch installation site. Use “pip” as the package, “stable” for the build, and everything else should match your software versions. Open up the Windows command prompt (search “cmd” in the start menu) and paste in the commands. https://pytorch.org/get-started/locally/

If it tells you “x is not a supported wheel on this platform”, make absolutely sure your options match and that your Python installation is 64-bit, or try different versions of Python or CUDA.

While you’re there, paste in this (pip3, not pip): pip3 install numpy opencv-python

Step 3: Download https://github.com/xinntao/ESRGAN along with the two pretrained models.

Step 3.5 (READ IF YOU HAVE AN AMD/INTEL GPU): Go into test.py and change “device = torch.device(‘cuda’)“ to “ device = torch.device(‘cpu’)”.

Step 4: Follow the instructions in ESRGAN’s “Quick Test” section. I highly recommend running the network interpolation command, because the default, full monty SRGAN model can make images a bit noisy whereas PSNR is too blurry. I usually use 0.8.

Ignore any “nn.Upsampling is deprecated” errors, they have no effect on the results. Do not use PNGs with alpha channels or you will get an “index out of range” error.k

Make sure to include the model path, or else you will get an “IndexError: list index out of range“ error. Either navigate to the ESRGAN directory via the command prompt, or save the commands you want as a .bat file in the ESRGAN directory.

.

Optional Step 5, for SFTGAN: Download SFTGAN here and follow the PyTorch instructions under “Test Models”. Make sure to download the pretrained models: https://github.com/xinntao/SFTGAN

SFTGAN is similar to ESRGAN. It tries to make more intelligent guesses about what kind of details it’s trying to retrieve, and it looks better in some cases (outdoor photographs, textures) and worse in others. The requirements are the same as ESRGAN.

Just Cause 2 texture comparison: https://imgur.com/a/JDsvVei

Place the images for  SFTGAN in data/samples. Unlike ESRGAN, you need to scale up the images by 4x before you run it. SFTGAN will downscale the image, then upscale it again, leaving the resolution unchanged. I use nearest neighbor scaling, but cubic scaling could be worth a try if dithering’s a problem.

Additional Info: I’ve been seeing this tutorial on other sites now and there’s a couple of things I need to clear up:

1. ESRGAN’s default networks were trained on a very specific type of image. That is, high resolution photographs downscaled with bicubic sampling. So it struggles with pics that went through color reduction or heavy dithering (like you see with old prerendered sprites), pictures that are blurry (I’d recommend downscaling them by 50%, as you’ll still get a 2x resolution boost), and pixel art. It may be possible to “tune” it by training it on specific categories of images. I’ve had some success adding small amounts of noise to the images to help it out, or blurring out color banding or sharp edges.

Other tips here:  https://boards.fireden.net/v/thread/443447437/#443455703

2. ESRGAN and SFTGAN are not the only super-resolution programs out there. The former is a modification of the existing SRGAN program from 2 years ago, and there’s new ones coming out every week. Not all of them exist as open source programs and not all of them run on Windows (Ubuntu 16.04 seems to be the “standard” OS for neural network programs), but I recommend you give them a try too, since there’s no one best way of upscaling photos. You could even shell out cash for Let’s Enhance or AI Gigapixel if you have more money than time.

3. Waifu2x is (as far as I know) the state of the art for illustrations and it’s what I’d recommend for HD anime reaction images.

Белые клетки крови, в том числе и Т-клетки, должны реагировать , когда организм борется с вторгающимися патогенами. Исследователи из Института Солка недавно получили изображения пакетированых критических рецепторов, присутствующих на поверхности Т-клеток при активации.

Солка ученые использовали свет листов супер-разрешением изображений, чтобы захватить перегруппировку рецепторов Т-клеток из…

Get PDF version of this post

Microscopy has been used for hundreds of years to decipher biological mechanisms. The advent of fluorescence microscopy and the ability to label specific proteins as well as to image live specimens allowed for even more discoveries. Boundaries have been pushed further by the development of super-resolution microscopy which awarded a Nobel Prize in Chemistry to Eric Betzig, Stefan Hell and William Moerner. It still bears notable problems that Eric Betzig cites in his new paper: fixed samples, phototoxicity and the need of an enormous photon budget. To bypass these limitations, they have been using SIM (Structured Illumination Microscopy) with the highest numerical aperture possible, and this way enhancing resolution while limiting the excitation to only of small portion of the cell volume and reducing phototoxicity. To achieve even higher resolution, patterned activation of photoswitchable fluorophores was used, allowing for further resolution enhancement and faster acquisition times. Still images convey the vast improvements over conventional optical microscopes but impressive videos can be seen on the online version.

Above : Two approaches for improved live-cell imaging at sub-100-nm resolution. (Left) Association of cortical actin (purple) with clathrin-coated pits (green), the latter seen as rings (inset) at 84-nm resolution via a combination of total internal reflection fluorescence and structured illumination microscopy at ultrahigh numerical aperture (high-NA TIRF-SIM). (Right) Progression of resolution improvement across the actin cytoskeleton of a COS-7 cell, from conventional, diffraction-limited TIRF (220-nm resolution), to TIRF-SIM (97-nm resolution), and nonlinear SIM based on the patterned activation of a reversibly photoswitchable fluorescent protein (PA NL-SIM, 62 nm resolution). Scale bars, 2 μm (left); 3 μm (right).