- Видео 67
- Просмотров 495 656
Duke Computational Optics Lab
Добавлен 23 ноя 2016
Hello - I'm Dr. Roarke Horstmeyer and I lead the Computational Optics Lab at Duke University. I work on new ways to capture and process high-resolution images, with a particular interest in biomedical applications. This channel shares some videos related to our research. Hopefully, this sparks some interest in what we do!
Please also check out some of my fun microscopy videos on Tiktok: www.tiktok.com/@dukemicroscopes
More details about my research are available at deepimaging.io and horstmeyer.pratt.duke.edu.
Please also check out some of my fun microscopy videos on Tiktok: www.tiktok.com/@dukemicroscopes
More details about my research are available at deepimaging.io and horstmeyer.pratt.duke.edu.
Rapid 3D whole-slide microscopy with a multi-camera array scanner MCAS)
Optical microscopy has long been a standard technique for clinical diagnosis. Increasingly, machine learning and AI software is now being adopted to assist with such diagnostic efforts. Unfortunately, however, today's digital microscopes are not effective at imaging large specimens in 3D. This limitation is particularly felt in the field of cytology, where patient specimens are both spread over large areas and are thick. Here, we present a new computational microscope that is optimized to rapidly scan such thick cytology slides in 3D. Our new Multi-Camera Array Scanner (MCAS) contains 48 compact microscopes closely arranged to scan large areas in parallel - 54x72 mm2, large enough to fit ...
Просмотров: 339
Видео
What is diffuse correlation spectroscopy?
Просмотров 1222 месяца назад
In this short lecture, Melissa Wu, a PhD student in the Computational Optics Lab at Duke University, explains diffuse correlation spectroscopy (DCS). This noninvasive optical method has been used within biomedical imaging, biophotonics and optical sensing to detect blood flow within tissue for several decades. Recent advances in new optical sources, single-photon detectors and data processing m...
Machine Learning and Imaging Lecture 12: Convolutional Neural Network Implementation Details
Просмотров 1192 месяца назад
In this lecture, Dr. Horstmeyer provides an overview of convolutional neural networks (CNN's) from a practical implementation standpoint. He covers the various features needed to get a CNN up and running in Tensorflow, as well as the primary options and variables that can be tuned to successfully train and test a CNN. Additional details can be found at deepimaging.github.io/ #machinelearning #c...
Machine Learning and Imaging Lecture 11: Back-propagation and auto-differentiation software
Просмотров 803 месяца назад
This lecture explains the key computational methods that underpin the world's current ability to efficiently train machine learning and deep learning models. These "auto-differentiation" software methods allow us to efficiently perform the key step of "back-propagation" in algorithm training, wherein the weights of a neural network are efficiently updated to make the algorithm "smarter", i.e. p...
Machine Learning and Imaging Lecture 10, Part 3: Ingredients of a Convolutional Neural Network
Просмотров 2893 месяца назад
In this lecture, Dr. Horstmeyer finishes describing the basic components of a convolutional neural network (CNN). This includes optimization methods to improve CNN performance, such as the use of batch normalization, dropout, data augmentation, and methods to initialize network weights. Additional resources are available at deepimaging.github.io #machinelearning #cameras #medicalimaging #ai
Machine Learning and Imaging Lecture 10, Part 2: Ingredients of a Convolutional Neural Network
Просмотров 2103 месяца назад
In this lecture, Dr. Horstmeyer continues to describe the basic components of a convolutional neural network (CNN). He analyzes the dimensionality of data and weights within each layer, presents details regarding convolutional strides and pooling operations, and overviews the various non-linear functions available for weight activation. He additionally discusses regularization strategies and pr...
Machine Learning and Imaging Lecture 10, Part I: Ingredients of a Convolutional Neural Network
Просмотров 2333 месяца назад
In this lecture, Dr. Horstmeyer outlines the basic components of a convolutional neural network (CNN). By focusing on the initial task of image classification, Dr. Horstmeyer mathematically defines the input and outputs of a CNN classifier, the operations of the convolutional layers composed of optimized weights, the non-linear and pooling operations that are often used, as well as additional k...
Machine Learning and Imaging Lecture 9: “Deep” Networks - Theoretical Motivations
Просмотров 1564 месяца назад
Machine Learning and Imaging Lecture 9: “Deep” Networks - Theoretical Motivations
Machine Learning and Imaging Lecture 7-8: Linear and Logistic Classification
Просмотров 604 месяца назад
Machine Learning and Imaging Lecture 7-8: Linear and Logistic Classification
Computational 3D topographic microscopy from terabytes of data per sample
Просмотров 5645 месяцев назад
Computational 3D topographic microscopy from terabytes of data per sample
Machine Learning and Imaging Lecture 6: Ingredients for Machine Learning
Просмотров 1235 месяцев назад
Machine Learning and Imaging Lecture 6: Ingredients for Machine Learning
Machine Learning and Imaging Lecture 5: Introduction to Optimization
Просмотров 966 месяцев назад
Machine Learning and Imaging Lecture 5: Introduction to Optimization
Machine Learning and Imaging Lecture 4: Mathematical Preliminaries for Discrete Functions
Просмотров 1526 месяцев назад
Machine Learning and Imaging Lecture 4: Mathematical Preliminaries for Discrete Functions
Machine Learning and Imaging Lecture 3: Fourier Transforms in imaging
Просмотров 1716 месяцев назад
Machine Learning and Imaging Lecture 3: Fourier Transforms in imaging
Machine Learning and Imaging Lecture 2: Signal Processing Preliminaries
Просмотров 776 месяцев назад
Machine Learning and Imaging Lecture 2: Signal Processing Preliminaries
Tensorial Tomographic Fourier Ptychography - a new type of 3D microscope
Просмотров 2947 месяцев назад
Tensorial Tomographic Fourier Ptychography - a new type of 3D microscope
3D imaging of cultural artifacts with the multi-camera array microscope
Просмотров 1837 месяцев назад
3D imaging of cultural artifacts with the multi-camera array microscope
Machine Learning and Imaging Lecture 1: Mathematical Preliminaries for Continuous Functions
Просмотров 2028 месяцев назад
Machine Learning and Imaging Lecture 1: Mathematical Preliminaries for Continuous Functions
Machine Learning and Imaging Lecture 0: Introduction
Просмотров 5118 месяцев назад
Machine Learning and Imaging Lecture 0: Introduction
Deep blood flow extraction via diffuse correlation spectroscopy with a SPAD array
Просмотров 2798 месяцев назад
Deep blood flow extraction via diffuse correlation spectroscopy with a SPAD array
Identifying immune cell types with label-free multi-photon imaging and deep learning
Просмотров 2078 месяцев назад
Identifying immune cell types with label-free multi-photon imaging and deep learning
High-resolution 3D video with a Fourier light field camera array
Просмотров 1,2 тыс.8 месяцев назад
High-resolution 3D video with a Fourier light field camera array
Video of cellular dynamics captured across 96 well-plate with a multi-camera array microscope - full
Просмотров 979 месяцев назад
Video of cellular dynamics captured across 96 well-plate with a multi-camera array microscope - full
Video of cellular dynamics captured across a 96 well-plate with a multi-camera array microscope
Просмотров 2349 месяцев назад
Video of cellular dynamics captured across a 96 well-plate with a multi-camera array microscope
Dynamic motion of hundreds of zebrafish embryo revealed by computational microscope
Просмотров 198Год назад
Dynamic motion of hundreds of zebrafish embryo revealed by computational microscope
Multi-gigapixel 3D image of an oil painting
Просмотров 946Год назад
Multi-gigapixel 3D image of an oil painting
Automated pose and eye tracking with the multi-camera array microscope
Просмотров 279Год назад
Automated pose and eye tracking with the multi-camera array microscope
Imaging Across Multiple Spatial Scales with the Micro Camera Array Microscope
Просмотров 704Год назад
Imaging Across Multiple Spatial Scales with the Micro Camera Array Microscope
Automated zebrafish pose estimation and eye tracking
Просмотров 63Год назад
Automated zebrafish pose estimation and eye tracking
Introduction to Fourier ptychography
Просмотров 4,6 тыс.Год назад
Introduction to Fourier ptychography
Im convinced this man is actually a lizard person
What resolution does a SPAD sensor have?
Great question! One of the newer ones in our lab has 500x500 pixels with a 16 micrometer pixel size, and there are now commercially available SPAD arrays with up to 3 Megapixels in them. With a properly selected lens, they can get pretty great resolution!
This is really cool! Roarke's videos are just getting more awesome by the day!
Thanks!!
Im suddenly watching this in 2024 🤣
Woah
Still doesn't make sense
It’s so odd when you actually think about it.
Well it's good to know what the creator of salad fingers is up to nowadays.
Funny how you explained the entire process without giving credit to the biggest contributer of modern cameras. An arabic scientist who came up with this. You said 18th century but this was before 14th century
Crystals... Yes Crystals and minerals are earths magic. Phones use crystals wtf
Casper Mission
witch craft i say
Sir do cameras we use today weather dslr or cameras in phones also has rolling shutter pls reply
Nearly all phones and DLSR cameras use a rolling shutter sensor. There are some exceptions, but most global shutter cameras are used within more scientific applications
This is a question that I had and when I asked this question, the answer I got was "funding reasons" - now I understand thanks to this video!
Glad I could help!
Roarke - this is Vivek! really love this - so cool and so much better than doing those tedious safety orientations! :)
Thanks glad I could help!
Those rodents have the best healthcare in the USA. Only animal with Universal Healthcare.
idc how manytimes this shit is explained to me its fkn magic
Man just say black magic....
Sir in rolling shutter you said that image is not capturing at same time but sir in video also as we know images are their so sir video is used as evidence so sir anyone can say it is fake image because it is not capturing at same moment it will loose integrity of images
Good question! Yes, rolling shutter cameras (for both images and video) do not capture everything at the same moment - that doesn't necessarily mean the image is "fake", but it does mean that it will not produce the same data as a global shutter camera. Artifacts can and will appear if things in the scene are moving, which you must be careful about during interpretation
@@dukemicroscopes sir it is producing fake images only because first upper part is formed then lower part and if suppose I gave any reaction then instantly I give another reaction but camera will captured mixture of two reaction which will be fake so sir it is losing integrity of images and anyone can say that it is fake
@@dukemicroscopes sir pls answer
Dear Melisaa! Is probe design is effective for measurement of Crebral blood flow ?
Yes, we have measured cerebral blood flow with our probe! Please see our recent paper about this here: arxiv.org/abs/2403.03968
@@dukemicroscopes thank you.. im doing Msc in biomedical engineering and working on thesis about DCS probe design. You’re research is the most recent and very helpful. Best regards!
They come along way ! That light picking up every single pore on my face
Let me ask everyone a question. We can take photos and videos on a phone. Similarly, you can copy the photo on the phone to any paper, but why can't you copy the video on the phone to a paper? Shouldn't Sherry play the video like this on the paper?
What's the point of this video? I can't make out a single brush stroke, the structure of the canvas or anything else that makes this a painting. The way you present it, it's just a meaningless heightmap with a graphics texture.
They’re not in the inside there on the outside
No u didn't really explain how digital camera works😢
This is so cool!!!! 😎
Great to see instantly applicable research!
Please post a link to demo software.
Obrigado! Assisti dezenas de vídeos e nenhum explicava como os aparelhos ópticos giravam a imagem invertida. Deus te abençoe!
This is an excellent video about DCS, Melissa. Thanks for your sharing!!
Dr. Roarke,Great research!! by the way, if you need any optical equipments, please contact me, thank you.
Is there any open-source code FPM image processing?
Amazing Inspiring
Nice presentation on interesting topic! The explanation of the technology and its application was clear and concise. Ten minutes video length was about right. Thank you for sharing.
Thank you so much, professor! I have been waiting for something like this from you. Really helpful 🙂
brro sir ji you are so underrated
Im watching this at 3:40 am. all of a sudden, i needed to know how this thing worked. 😂😂😂
Nice
Respect
Nice
nice
Makes me wonder how many people were stoned n was like “how tf does a camera work”
Nice explanation about camera how it's work thanks for making this video
if we can somehow figure out how to stop a moment in time and keep it, what else could we do? seriously....
Its witchcraft I tell ya!!
I still dont get it🤷
Hi
Why did he smack him 😂
As someone who owns a microscope and works at a actual lab, this is peak comedy.
Tease.