Christina Bornberg

Deep learning for ophthalmology 👁👩🏻‍💻 | datascEYEnce! column in the Computer Vision News @RSIPVision | my main account: @variint

[Christina Bornberg @datasceyence]

Book club topics of this week were:
- Growing Neural Cellular Automata: https://t.co/6FVr9DFna3
- HyperNCA: https://t.co/MZOnJ6vd3u
- Medical: Retinoblastoma (cancer in the eye, almost exclusively found in young children)

[Christina Bornberg @datasceyence]

Another deep learning for ophthalmology story is online in the December version of the @RSIPvision News! 👀👀 This time, I interviewed @ignaciorlando on how to go the extra mile for creating a product out of research! You can read the full story here:
https://t.co/6IsCKpwWWi

[INSIGHTeyehub @INSIGHTeyehub]

📢👁️Applications are open for a funded PhD Studentship in Artificial Intelligence for Ocular Imaging @UCLeye @UCLBiochemEng1 @MoorfieldsBRC

⏰Deadline: 8 Jan 2024
https://t.co/YjuLEZ98xk
#ArtificialIntelligence #DataScience

[Christina Bornberg @datasceyence]

Also, the preprint is available on arxiv: https://t.co/ICcp6m6pBq and you can find the MICCAI 2023 version here: https://t.co/SezOvUTZyS

[Christina Bornberg @datasceyence]

For the November edition of the @RSIPvision computer vision magazine, I interviewed Robbie Holland from @BioMedIAICL about his work on temporal biomarker detection for age-related macular degeneration in OCT scans: https://t.co/iiWnItJ64Y

[Christina Bornberg @datasceyence]

Also, to make it possible for other researchers in the field to work with their pre-trained foundation model, they made the code as well as weight files available on GitHub: https://t.co/bazMIdF456

[Christina Bornberg @datasceyence]

Another month, another datascEYEnce highlight! For the October version of the @RSIPvision Computer Vision Magazine, I interviewed @Yukunzhou19 about RETFound - a foundation model for retinal images: https://t.co/4r0wDFHwYG

[Christina Bornberg @datasceyence]

Their paper recently got published in nature! You can find the full publication here: https://t.co/XRWsjTGVvN

[Moritz Fuchs @mofuchs1]

Huge thanks to @RSIPvision and @variint for featuring our synthetic cataract surgery research!

Thanks to @softwarecampus1, Antoine Sanner, @anirbanakash, and many more!
@MiccaiStudents take a look here:
📜https://t.co/7fEw0BEJAb
🧩https://t.co/UG0EJFNhbQ

[Yannik Frisch @YannikFrisch]

Huge thanks to @RSIPvision and @variint for featuring our research paper on synthetic cataract surgery 💻👁️
Honored to see our work recognized and shared with a broader audience.

Dive into the article here: https://t.co/DBkUlVSQsF

[Christina Bornberg @datasceyence]

His publication "Synthesising Rare Cataract Surgery Samples with Guided Diffusion Models" will soon appear at #MICCAI2023! Until then, you can find the arXiv version here: https://t.co/GSd7r4vciY and you can find their code on GitHub: https://t.co/dhgOqsZUc9

[Christina Bornberg @datasceyence]

Hello everyone! This month, I am introducing you to @YannikFrisch and his work on diffusion models applied to cataract surgery videos! Read the whole story here: https://t.co/cbHZQ1fEWs

[Christina Bornberg @datasceyence]

News from the deep learning for ophthalmology field - RETFound just got published in Nature! In a self-supervised phase, a good representation of the eye is learned and can then be fine-tuned with a smaller labelled dataset for disease detection!

[Pearse Keane @pearsekeane]

1/

🚨🚨 New paper alert 🚨🚨

Introducing RETFound, a foundation model for ophthalmology

We’re super excited about this and hope it will act as a #Cornerstone for global efforts to prevent blindness through #AI

@UCLeye @Moorfields

#OpenAccess @Nature
https://t.co/DsvKgeUH5m

[Christina Bornberg @datasceyence]

Also, in case you are attending #MICCAI2023, watch out for his newest research project "Self-supervised learning via inter-modal reconstruction and feature projection networks for label-efficient 3D-to-2D segmentation" - https://t.co/cnCz7S2mAu

[Christina Bornberg @datasceyence]

Hello #ophthotwitter! You can read the first interview of the datascEYEnce column in Computer Vision News by @RSIPvision now!! The topic is a recent publication by José Morano Sánchez on multiple instance learning applied to AMD lesions in fundus images: https://t.co/n1Uj0h0Yh0

[Christina Bornberg @datasceyence]

On page 27 of the Computer Vision News, you can find some more implementation details!

[Christina Bornberg @datasceyence]

Deep learning session intro in the lab!! We covered PyTorch datasets, samplers and dataloaders!

[Christina Bornberg @datasceyence]

More detailed information can be found in the paper: https://t.co/SNBYiav3u4 and their code is available here: https://t.co/1ZC0zPBSRf

[Christina Bornberg @datasceyence]

"Beyond the Pixel-Wise Loss for Topology-Aware Delineation", published at CVPR 2018 by a research team bases in Switzerland, focuses on the task of delineation - the action of indicating the exact position of a border or boundary.

[Christina Bornberg @datasceyence]

Due to the large impact on the topology when only a few pixels are misclassified, they introduce a penalty term that considers the number of connected components and number of holes. Additionally, they enhance their method through the introduction of "recursive refinement."