Dimitrije Antić

Most research time isn’t spent thinking of ideas. It’s spent checking prior work, testing hypotheses, debugging silent bugs, organizing results, making slides. The parts where weeks of work silently disappear. I tried encoding some of that experience as “Agent Skills”.

me: "can you use whatever resources you like, and python, to generate a short 'youtube poop' video and render it using ffmpeg ? can you put more of a personal spin on it? it should express what it's like to be a LLM" claude opus 4.6:

Do not execute yet. Ask clarifying questions until we align and then make a plan on which I sign off.

Novak Djokovic.

When AI automates parts of your job, it increases the value of the parts it doesn't. I.e., it makes you more valuable.

If you're stuck, simplify. If you're still stuck, simplify again.

One of these men was banned from Australia.

Pre-PhD: I have a question Mid-PhD: I have a headache

@Vinc3nt_Leroy @ICCVConference I believe this is the correct: media.eventhosts.cc/Conferences/IC…

Hey @ICCVConference which poster number is right for the session? The book and the website are not consistent

If you want to learn more about general models for 3D and semantic understanding, or just have a quick chat, come meet us this afternoon at posters #462 (HAMSt3R) and #540 (PanSt3R) :)

#SDFit will be presented at #ICCV2025: ⏰ 21 Oct, 15:00-17:00 📍 Poster #892 🌐 Webpage: anticdimi.github.io/sdfit 👨‍💻 Code: github.com/anticdimi/sdfit Great collaboration between: 👉 @UvA_Amsterdam → me, @gkpaschalidis, @theogevers, @dimtzionas 👉 @MPI_IS → @saidwivedi, @sha2nk_t Thanks to reviewers & ACs for feedback! (6/6🧵)

To conclude: 👉#SDFit exploits an explicit, morphable 3D shape prior (mSDF) to generate 3D hypotheses, and a feedback loop to fit the mSDF hypothesis to an img 👉#SDFit is a good step toward robust 3D perception, and provides insights toward methods that better 'understand' rather than 'hallucinate' (5/6🧵)

Intelligent systems need to understand #3D Objects from single images. However, recent SotA methods work well for unoccluded views & common poses, and prioritize shape over pose. We tackle this by developing #SDFit #ICCV2025 a novel Render-n-Compare framework for 3D objects that: 👉 Is robust to occlusion & uncommon poses 👉 Treats both Shape & Pose as 1st class citizens (1/6🧵)

depth ambiguity

Stochastic Gradient Descent

📢 Do u want to #Review for #CVPR2026, but u might not be in the reviewer pool? 📢 Do u also have prior publications at venues like #CVPR, #ECCV, #ICCV? 👉 Please DM me (or other ACs) with your: - First name - Last name - Email - OpenReview profile (plz update it with full info) - GoogleScholar profile 🆘 Deadline 26/9

Software engineers shouldn't fear being replaced by AI. They should fear being asked to maintain the sprawling mess of AI-generated legacy code their employer's systems will soon run on. Because that one will actually happen.