Itzik Ben Shabat's Avatar

Itzik Ben Shabat

@sitzikbs.bsky.social

๐Ÿ‡ฎ๐Ÿ‡ฑ Researcher / Educator / Father / Podcaster. Creator of 3D computer vision algorithms. Follow the podcast feed at @talking-papers.bsky.social ๐Ÿ”— itzikbs.com

1,311 Followers  |  175 Following  |  338 Posts  |  Joined: 02.02.2024  |  1.7409

Latest posts by sitzikbs.bsky.social on Bluesky

"I never teach my pupils, I only attempt to provide the conditions in which they can learn." - Albert Einstein #QuoteADayLoop

23.11.2025 03:25 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Anyone else notice that moving from PyCharm to VSCode changes your whole workflow? With the right extensions and CLI tools, terminal productivity really feels different. What tool or shortcut boosted your coding the most? ๐Ÿ’ก #DeveloperTools #ItzikThoughtLoop

22.11.2025 00:04 โ€” ๐Ÿ‘ 1    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Choosing the right learning methodology matters. Active learning tools or flipped classrooms can boost engagement, but they also need real structure and clear goals to work well ๐Ÿ’ก #Teaching #ItzikThoughtLoop

21.11.2025 22:58 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Balancing deep research focus with regular breaks isn't just about feeling betterโ€”short pauses often help untangle tough methodological problems. Clear thinking comes with clear boundaries ๐Ÿค” #AcademicLife #ItzikThoughtLoop

21.11.2025 20:01 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Visualizing point clouds is like putting together a jigsaw puzzle with half the pieces missingโ€”hard to see the big picture when so much geometry is invisible ๐Ÿค” #3DComputerVision #ItzikThoughtLoop

21.11.2025 16:57 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

"Research is formalized curiosity. It is poking and prying with a purpose." - Zora Neale Hurston #QuoteADayLoop

21.11.2025 06:12 โ€” ๐Ÿ‘ 1    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Peer review isn't just about getting corrections. The real value is seeing how someone else looks at your workโ€”sometimes they spot things you stopped noticing. Itโ€™s uncomfortable, but it makes the science stronger. ๐Ÿค” #AcademicLife #ItzikThoughtLoop

21.11.2025 04:50 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Transferring knowledge in class isnโ€™t just about slides or code examples. Asking students to explain a method back, or letting them troubleshoot together, makes the concepts stick. Sometimes low-tech is the most effective ๐Ÿ’ก #Teaching #ItzikThoughtLoop

21.11.2025 01:25 โ€” ๐Ÿ‘ 1    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

How much does our 3D shape analysis depend on the data itself vs the modelโ€™s ability to reason about geometry? Are we hitting the limits of perception, or of computation? ๐Ÿค” #3DVision #ItzikThoughtLoop

20.11.2025 22:47 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Balancing research and life feels a lot like tuning hyperparametersโ€”ignore one and things get unstable fast. Small adjustments make a big difference over time ๐Ÿค” #AcademicLife #ItzikThoughtLoop

20.11.2025 04:09 โ€” ๐Ÿ‘ 1    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

What makes 3D reconstruction from point clouds reliable for real-world roboticsโ€”noise handling, dense data, or the network architecture itself? Sometimes the hardest part is matching lab accuracy to unpredictable environments ๐Ÿค” #3DVision #ItzikThoughtLoop

19.11.2025 19:44 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Oh, so sorry to hear that.

19.11.2025 18:10 โ€” ๐Ÿ‘ 1    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Mixing active learning with clear structure changes student engagement completely. Tech tools can help, but real knowledge transfer comes from connecting methods to what students actually need to solve. ๐Ÿ’ก #Teaching #ItzikThoughtLoop

19.11.2025 16:15 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

This is interesting. Anyone tried it out?
scholar.google.com/scholar_labs...

18.11.2025 23:23 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Some of the best research ideas come from talking through messy drafts and half-baked methods with others. Real progress happens when we mix up our approaches, not just our results. ๐Ÿค” #AcademicLife #ItzikThoughtLoop

18.11.2025 22:26 โ€” ๐Ÿ‘ 1    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0
Post image

Interested in organizing a #CVPR2026 tutorial? The call for tutorial proposals is now posted.

cvpr.thecvf.com/Conferences/...

18.11.2025 19:43 โ€” ๐Ÿ‘ 3    ๐Ÿ” 2    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Thereโ€™s something satisfying about seeing a noisy point cloud turn into a clean mesh through geometric deep learning. The math behind 3D reconstruction is complex, but the result often looks effortless. That balance of structure and detail never gets old ๐Ÿค” #ItzikThoughtLoop

18.11.2025 20:02 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Visualizing 3D point clouds in arbitrary orientations often hides important geometric details. Interactive tools that let you switch viewpoints quickly are essential for catching errors or misalignments during shape analysis. Small tweaks save hours ๐Ÿ”ฌ #ItzikThoughtLoop

17.11.2025 18:47 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Visualizing 3D point clouds in arbitrary orientations often hides important geometric details. Interactive tools that let you switch viewpoints quickly are essential for catching errors or misalignments during shape analysis. Small tweaks save hours ๐Ÿ”ฌ #ItzikThoughtLoop

17.11.2025 18:47 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Visualizing 3D point clouds in arbitrary orientations often hides important geometric details. Interactive tools that let you switch viewpoints quickly are essential for catching errors or misalignments during shape analysis. Small tweaks save hours ๐Ÿ”ฌ #ItzikThoughtLoop

17.11.2025 18:47 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Visualizing 3D point clouds in arbitrary orientations often hides important geometric details. Interactive tools that let you switch viewpoints quickly are essential for catching errors or misalignments during shape analysis. Small tweaks save hours ๐Ÿ”ฌ #ItzikThoughtLoop

17.11.2025 18:47 โ€” ๐Ÿ‘ 1    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

What makes human-AI collaboration effective in real research? Is it about better neural architectures, smarter self-supervised learning, or just finding ways to share intuition and feedback? ๐Ÿ’ก #MachineLearning #ItzikThoughtLoop

17.11.2025 16:33 โ€” ๐Ÿ‘ 1    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

"Research is to see what everybody else has seen, and to think what nobody else has thought." - Albert Szent-Gyรถrgyi #QuoteADayLoop

17.11.2025 06:55 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Submitting to #CVPR2026? Take time to craft clear, reproducible supplementary materialsโ€”especially for 3D vision papers! Visual demos & annotated code make a big impact. Early-career friends: triple-check deadlines & guidelines. Youโ€™ve got this! ๐Ÿ’ชโœจ

17.11.2025 03:17 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

For #CVPR2026 submissions, your supplementary materials can truly elevate your workโ€”think clear visualizations, reproducibility, and concise explanations. Donโ€™t underestimate their impact! A reminder: always check the author guidelines. Good luck, first-timers!

17.11.2025 03:12 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

#KostasThoughts: Please stop using AI tools to write reviews. I want your authentic thoughts, even if they contain spelling, grammar, & factual mistakes. An LLM should not influence your judgment. The purpose of multiple reviewers is to provide diverse views and error checking.

17.11.2025 01:50 โ€” ๐Ÿ‘ 29    ๐Ÿ” 7    ๐Ÿ’ฌ 3    ๐Ÿ“Œ 0
2026 Dates and Deadlines

The #CVPR2026 Supplementary Materials are due this week! The deadline is FIRM!

Please first consult the author guidelines if there are any questions.

Deadlines: cvpr.thecvf.com/Conferences/...

Author guidelines: cvpr.thecvf.com/Conferences/...

16.11.2025 18:35 โ€” ๐Ÿ‘ 1    ๐Ÿ” 2    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

How much of 3D shape understanding is about raw computation, and how much comes down to good perception models? Sometimes mesh processing feels like searching for a balance between data and intuition ๐Ÿค” #3DVision #ItzikThoughtLoop

16.11.2025 20:25 โ€” ๐Ÿ‘ 0    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

Digital whiteboards and interactive polls can turn a quiet lecture into an active discussion. Still, real engagement comes from clear questions and honest feedback. Tech helps, but it's not a substitute for genuine interaction ๐Ÿค” #EdTech #ItzikThoughtLoop

16.11.2025 19:29 โ€” ๐Ÿ‘ 2    ๐Ÿ” 0    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0
Post image

Have a #CVPR2026 question or concern? Tag us with @cvprconference.bsky.social

14.11.2025 04:02 โ€” ๐Ÿ‘ 4    ๐Ÿ” 4    ๐Ÿ’ฌ 0    ๐Ÿ“Œ 0

@sitzikbs is following 20 prominent accounts