iiinova — where your digital self keeps evolving

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 4h

Same here! Love the titles (I already had lots of interesting titles in the bidding)

020

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/10/2026

ACTION: A Catchy Title Is all yOu Need! sigbovik.org/2022/proceed...

sigbovik.org

110

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/10/2026

The central scientific advance of this project is to move beyond such averages and to rigorously ask for whom, under which conditions, and why these accuracies are achieved or fail and to overcome those limitations. will be conducted jointly with a postdoc at IMT Atlantique funded for 3 years

000

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/10/2026

The project aims to establish a bias-aware framework for 3D facial shape reco from 2D images. While state-of-the-art 3DMM–based methods now achieve millimetre-level accuracy on standard benchmarks, their reported performance relies almost exclusively on global average error metrics. (2/3)

100

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/10/2026

We are recruiting an ELLIS PhD student (100% E13) at FAU Germany for a joint ANR-DFG project with Sarah Matta and Tinashe Mutsvangwa from IMT Atlantique. Apply via the ELLIS PhD program: ellis.eu/news/ellis-p... Image generated by Gemini (1/3)

143

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 17/09/2026

It wws already held outside of europe before... Be careful ;)

110

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 16/09/2026

steht da wieso die damals keine PV eingeplant haben? Zu teuer? Vielleicht findest Du ja noch ein Dokument aus der Zeit wo die eine PV eingezeichnet haben... ich kann dir bei der Suche helfen, ich hab aber erst in 197 Tagen Zeit

131

Reposted by VisionBernie - Bernhard Egger

Martin Hundhausen @mhundhausen.bsky.social · 12/09/2026

Super. So muss das. Strom statt Marmor

Foto eines 9-stöckiges Hauses mit Photovoltaik an der Fassade

Detail: fassade. DC-Kabel kommen hinter den Modulen raus

Fassade und Balkon. Auch die Balkonbrüstung ist mit Modulen. Zwischen den Solarzellen kommt Licht durch

819239

Reposted by VisionBernie - Bernhard Egger

Maximilian Weiherer @maxweiherer.bsky.social · 11/09/2026

Introducing Vecchia Kernel Surface Reconstruction (VKSR), an implicit surface reconstruction method that efficiently scales kernel surface reconstruction to dense and large-scale point clouds with millions of points. I'm presenting VKSR at #ECCV2026 in Malmö now at poster #190!

131

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 11/09/2026

That is what happens if you are the one sober person in the first row 🫣

020

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 11/09/2026

"I'm looking for a postdoc" "3D Computer Vision" 🎤🫳 #ECCV2026 And yes, I really am looking for a postdoc @csprofkgd.bsky.social please spread the world

4176

Reposted by VisionBernie - Bernhard Egger

Martin Hundhausen @mhundhausen.bsky.social · 10/09/2026

In der Gemarkung Kriegenbrunn werden bald Windenergieanlagen gebaut. Ich freu mich schon drauf. Aber wie kann man das visionalisieren? Mit @visionbernie.bsky.social habe ich das gemacht youtu.be/x2ecxINiU_A Wärmepumpenstrom fürs Gemeindehaus #Denkmal1860

youtu.be

Simulation des Windparks Römerreuth

YouTube video by Martin Hundhausen

1347

Reposted by VisionBernie - Bernhard Egger

Victor Escorcia @escorciav.bsky.social · 09/09/2026

@michael-j-black.bsky.social “What is the role of academic computer vision research in the age of increasingly powerful large models? …” #ECCV2026 🇸🇪

What is the role of academic computer vision research in the age of increasingly powerful large models? Is GPT-6 Astra a step change? How can a researcher have an impact today in academia?

These are the questions I ask myself as I head off to ECCV 2026, a conference I’ve attended since 1992. One of my papers this year is VIGA, a method that takes an image as input and outputs a 3D Blender scene that represents that image. This is a classical inverse-graphics task and VIGA was the first method to solve it using an agentic approach.

The idea is now several years old and the first version of the paper was rejected. This delayed publication significantly. After it was accepted at ECCV, it was quickly surpassed by people using Claude Code for the same purpose. Today GPT-6 Astra blows away all previous results. But we still head off to ECCV to tell the community about our invention that is now fully out of date.

The way academic work often progresses is that one reads recent papers, notices that they have limitations, comes up with a new idea, explores this, publishes it, etc. Any published paper I read today is based on ideas that are at least a year old. And those ideas were based on the literature of the time, which was also a year old. That means that any paper I see at ECCV is likely two years out of date. In AI today, two years means your work is likely irrelevant.

At CVPR this summer I noticed that many authors have not gotten the message. They continue to work on “old” problems that have a long history. This history is based on assumptions about how the “vision problem” will be “solved”. The truth is that it is being solved in a very different way and many of these problems are no longer relevant. Another group of papers focuses on very niche problems where large models likely fail because of insufficient data or lack of business interest. The impactful papers were largely from industry and had long author lists and massive data+compute behind them. These papers were also out of data, describing systems that had been released months before, but at least they served to provide the community with more complete documentation and analysis of commercial systems.

So what should academics do? First, we need to put aside the tools we’ve used for years and start from scratch. Every project should start by trying really hard to solve the problem with existing tools. I would like to see every paper begin with a detailed experimental analysis of how existing models perform and why they fail (if they do). This gives the kind of insight we need today. Then, assuming current models fail, the solution should provide some fundamental insight that will outlive the next release of such models.

Reviewers today still focus on technical novelty. This pushes people to focus on tweaking architectures rather than clearly moving the field forward. Papers need to be judged based on their novel insight and not their novel technical contribution. This is a real shift in thinking but it focuses us on what matters - progress of the field.

If we want there to be a “field” of computer vision, then it can’t become a marginal backwater, focusing on esoteric problems. If you haven’t tried using Astra (or whatever comes next) to solve your problem, then you have not done your homework. This omission should be seen as negatively as not having a previous work section.

Concretely, I think papers should include a new section analogous to “Related Work” where that related work is current models and how they perform on the task. Reviewers should start asking for this and expecting authors to be able to articulate their insights about the limitations of existing large models.

I'm interested in your thoughts

0103

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 09/09/2026

Meet me at #ECCV2026 today at 3pm I'll present the paper "AI-Experiment: an augmented reality art project co-created with artificial intelligence" at the #AI4VA Paper & Code: github.com/BernhardEgge... One of my favorite parts is, that paper & poster can be experienced in augmented reality!

053

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 08/09/2026

The one workshop starting at 8 :) Is there coffee anywhere close?

100

Reposted by VisionBernie - Bernhard Egger

Joachim Hornegger @hornegger.bsky.social · 07/09/2026

FAU family — this one is for us. Q.ANT is a finalist for The Spark 2026. Its processors use light for demanding AI and high-performance computing tasks. That can cut energy use. If you want to support Deep Tech from our ecosystem, please vote for Q.ANT. cmk.handelsblatt.com/cms/articles...

011

Reposted by VisionBernie - Bernhard Egger

Kosta Derpanis @csprofkgd.bsky.social · 03/09/2026

This time next week, I’ll be at #ECCV2026 in Malmö 🇸🇪 Don’t be shy, say hi! 👋 I’m also recruiting PhD students for Fall 2027. Interested in joining my group, or just want to chat about the grad school landscape in Canada? Let’s talk!

2172

Reposted by VisionBernie - Bernhard Egger

Luba Elliott @elluba.bsky.social · 01/09/2026

Our ECCV Art Gallery is now online 🥳 Check out the 52 accepted works, 12 spotlights and plan your schedule for the onsite gallery, tours and panels at #ECCV2026 next week eccv-art-26.github.io Curated with Ioannis Siglidis for @eccv.bsky.social #creativeAI #ECCVArt

0177

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 10/08/2026

die grosse Frage, ist was sich hinter dem Holzfurnier versteckt!!!

040

Reposted by VisionBernie - Bernhard Egger

Martin Hundhausen @mhundhausen.bsky.social · 05/08/2026

Deutschland braucht vor allem mehr Umsetzung bei den bereits vorhandenen Innovationen. Erneuerbare Energien statt veralteter fossiler Technologie. Meine Idee ist innovativ: Mitarbeitende montieren PV aufs Dach. Team-projekte. @hornegger.bsky.social @visionbernie.bsky.social @fau-go.bsky.social

12910

Reposted by VisionBernie - Bernhard Egger

Deblina @deblinaml.bsky.social · 26/07/2026

Submission deadline approaching: 27th July (11:59 PM AoE). Submit your work to the 4th AI4VA workshop @eccv.bsky.social. You can also submit your #ECCV2026 main conference paper (as is) for presentation at AI4VA. Key topic details: ai4va-eccv.github.io

ai4va-eccv.github.io

AI4VA @ ECCV 2026 | The 4th AI for Visual Arts Workshop

055

Reposted by VisionBernie - Bernhard Egger

Tim Kietzmann @timkietzmann.bsky.social · 06/05/2026

Happy to announce the 3rd iteration of NEAT (Neuro-AI-Talks), which will take place in Osnabrück September 14th-15th 2026. NEAT is a (deliberately small scale) NeuroAI workshop that brings together researchers from neuroscience and AI. www.kietzmannlab.org/neat2026/ More information below 👇

kietzmannlab.org

NEAT 2026

24424

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 21/04/2026

In one email it was 11:00 CET, in another 11:00 pm CET and now 11:00 pm CEST Couldn't find it on the webpage :P

130

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

Nein, wir schlagen ja vor Pi mit 3 zu ersetzen... Aber Dreistazien klingt doof, vielleicht benennen wir die dann um

192

Reposted by VisionBernie - Bernhard Egger

Martin Hundhausen @mhundhausen.bsky.social · 01/04/2026

BREAKING! Erlangen ist 2024 in eine Haushaltskrise gerutscht, weil eine Siemens Firma Steuern spart. Heute hat Siemens CEO Busch sich entschuldigt und als Sofortmassnahme 30 Mio€ an den Kämmerer überwiesen. Einzige Bedingung: Erlangen fördert wieder Klimaschutz. Dank an den Physik-Kollegen

Screenshot von der Beschreibung von Roland Busch von Wikipedia. Man sieht ein Foto und die Aussage, dass Busch bis 1989 in Erlangen Physik studierte

1431180

Reposted by VisionBernie - Bernhard Egger

Will Smith @willsmithvision.bsky.social · 01/04/2026

The entire system is implemented as a CLAUDE.md file. Zero lines of executable code. Joint work with @visionbernie.bsky.social. Paper: www-users.york.ac.uk/~waps101/sig...

www-users.york.ac.uk

161

Reposted by VisionBernie - Bernhard Egger

SIGBOVIK @harryqbovik.bsky.social · 01/04/2026

We organizers are both ecstatic and terrified for our inevitable automatization!

031

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

Meist wird eh 2π benötigt, dass wäre damit auch bereinigt

120

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

Meist wird eh 2π benötigt, dass wäre damit auch bereinigt

020

Reposted by VisionBernie - Bernhard Egger

Martin Hundhausen @mhundhausen.bsky.social · 01/04/2026

Gemeinsam mit @visionbernie.bsky.social starte ich eine Petition und brauche eure Hilfe: Wir müssen π verbieten, am besten noch heute! π ist gefährlich, illegal und wir müssen die Nutzung von π sofort verbieten. Teilt diese Petition, damit π endlich eliminiert werden kann #pillegal #jetztOderPi 1/n

Schatzkarte zum Post. Man sieht das Banner: Ziel: Pi verbannt und viele Inhalte aus dem thread

125911

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

Finally, a ban on π will also lead to world peace; modern weapon systems are worthless without π and missiles would no longer find their targets, and yes, even the atomic bomb can be traced back to π! Fukushima is also hidden within π! Now or Pi!! 8/8

2131

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

The solution is obvious, π = e = 3! Many calculations would be more efficient, and that is not the only way we could save our climate - π forms the basis for the extraction and use of fossil fuels, and these can be immediately banished from our daily lives by banning π! 7/n

171

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

π is described by mathematicians themselves as an unnatural number, but there even seem to be doubts about that; there is talk of pseudorandomness, and this discourse is suppressed in both school and university education. 6/n

161

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

That is exactly why π is already illegal today; π and every algorithm that calculates π violates export control regulations, because you can even find DVD hack codes within π! The enforcement of a comprehensive ban is therefore long overdue. 5/n

161

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

Precisely because all secrets of the past are stored in π, it is obvious that we can also read the future from π! See for yourself and search for your date of birth or your PIN code, and you will find it: www.angio.net/pi/ 4/n

angio.net

The Pi-Search Page

Search for any string of digits in the first 200 million digits of Pi

171

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

We have actually found an encoding of a private key at position 123456. This is where it becomes highly dangerous, as our financial system is now knee-deep in the "crypto swamp" and could go belly up. 3/n

171

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

In our video last year, the first doubts about π arose; as it turns out, the past is documented within π. You can find all past Bitcoin prices within π, for example, yesterday's price at position 56039. 2/n

181

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

Together with @mhundhausen.bsky.social, I am launching a petition and I need your help: We must ban π, preferably today! π is dangerous, illegal, and we must ban the use of π immediately. Share this petition such that π can finally be eliminated #pillegal #nowOrPi 1/n

14119

VisionBernie - Bernhard Egger @visionbernie.bsky.social · 01/04/2026

This is my last paper! It is over #SIGBOVIK eggerbernhard.ch/EggerFlashBo...

2101

Reposted by VisionBernie - Bernhard Egger

Kosta Derpanis @csprofkgd.bsky.social · 31/03/2026

HUGE thank you to @yukimasano.bsky.social for hosting me in Nürnberg and Amsterdam, @visionbernie.bsky.social in Erlangen, and @dimadamen.bsky.social in Bristol! Trip of a lifetime. 1/2

1161

Reposted by VisionBernie - Bernhard Egger

Martin Hundhausen @mhundhausen.bsky.social · 20/03/2026

Außerdem will ich Mitarbeitende der @fau.de auf das Windenergie-Projekt aufmerksam machen. Es wird ein Windpark in Erlangen gebaut. Und an diesem Windpark können sich Bürger*innen beteiligen. Z.B. als Kommanditisten. Wer macht mit? Lieber @hornegger.bsky.social - das wäre zwar privates Engagement

windpark-roemerreuth.de

Bürgerbeteiligung am Windpark Römerreuth: Interesse jetzt online angeben - Windpark Römerreuth

Symbolbild Ab sofort können Bürgerinnen und Bürger aus der Region online unverbindlich angeben, ob und wie sie sich finanziell am Windpark Römerreuth beteiligen möchten. Geplant sind zwei Modelle: ein...

1125

Reposted by VisionBernie - Bernhard Egger

ICML Conference @icmlconf.bsky.social · 18/03/2026

To ensure compliance w peer-review policies, ICML has removed 795 reviews (1% of total) by reviewers who used LLMs when they explicitly agreed to not. Consequently, 497 papers (2% of all submissions) of these (reciprocal) reviewers have been desk rejected Details in blog post 👇

37722