Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

FlashPortrait based on Wan2.1 14B: faster infinite-length portrait animation; - no identity drift; - tops Wan-Animate in FID/FVD scores over 5400+ frames.

Wildminder

3,909 subscribers

10,433 views • 8 months ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Another beast from Alibaba. Live Avatar: real-time, infinite-length streaming avatar generator; - 20 FPS, based on a 14B WanS2V trained with DMD. - locks identity; They plan to: add TTS, low VRAM, 3 steps, SVD quantz...

Wildminder

21,442 views • 8 months ago

China open-sourced a model that reconstructs any scene in 3D from a regular video, in real-time. one camera. no LiDAR. 10,000+ frames without falling apart. just walk around with your camera and watch the entire world get rebuilt in 3D at 20 fps. → runs at ~20 FPS on a single GPU → Stable over 10,000+ frames → Beats optimization-based methods on benchmarks → Works on drone footage, driving videos, indoor walkthroughs 100% open source.

China open-sourced a model that reconstructs any scene in 3D from a regular video, in real-time. one camera. no LiDAR. 10,000+ frames without falling apart. just walk around with your camera and watch the entire world get rebuilt in 3D at 20 fps. → runs at ~20 FPS on a single GPU → Stable over 10,000+ frames → Beats optimization-based methods on benchmarks → Works on drone footage, driving videos, indoor walkthroughs 100% open source.

Yasir Ai

253,764 views • 1 month ago

China open-sourced a model that reconstructs any scene in 3D from a regular video, in real-time. one camera. no LiDAR. 10,000+ frames without falling apart. just walk around with your camera and watch the entire world get rebuilt in 3D at 20 fps. → runs at ~20 FPS on a single GPU → Stable over 10,000+ frames → Beats optimization-based methods on benchmarks → Works on drone footage, driving videos, indoor walkthroughs 100% open source.

China open-sourced a model that reconstructs any scene in 3D from a regular video, in real-time. one camera. no LiDAR. 10,000+ frames without falling apart. just walk around with your camera and watch the entire world get rebuilt in 3D at 20 fps. → runs at ~20 FPS on a single GPU → Stable over 10,000+ frames → Beats optimization-based methods on benchmarks → Works on drone footage, driving videos, indoor walkthroughs 100% open source.

Superman

1,452,404 views • 1 month ago

MagicAnimate: Temporally Consistent Human Image Animation using Diffusion Model with Gradio demo local demo: This paper studies the human image animation task, which aims to generate a video of a certain reference identity following a particular motion sequence. Existing animation works typically employ the frame-warping technique to animate the reference image towards the target motion. Despite achieving reasonable results, these approaches face challenges in maintaining temporal consistency throughout the animation due to the lack of temporal modeling and poor preservation of reference identity. In this work, we introduce MagicAnimate, a diffusion-based framework that aims at enhancing temporal consistency, preserving reference image faithfully, and improving animation fidelity. To achieve this, we first develop a video diffusion model to encode temporal information. Second, to maintain the appearance coherence across frames, we introduce a novel appearance encoder to retain the intricate details of the reference image. Leveraging these two innovations, we further employ a simple video fusion technique to encourage smooth transitions for long video animation. Empirical results demonstrate the superiority of our method over baseline approaches on two benchmarks. Notably, our approach outperforms the strongest baseline by over 38% in terms of video fidelity on the challenging TikTok dancing dataset. Code and model will be made available.

MagicAnimate: Temporally Consistent Human Image Animation using Diffusion Model with Gradio demo local demo: This paper studies the human image animation task, which aims to generate a video of a certain reference identity following a particular motion sequence. Existing animation works typically employ the frame-warping technique to animate the reference image towards the target motion. Despite achieving reasonable results, these approaches face challenges in maintaining temporal consistency throughout the animation due to the lack of temporal modeling and poor preservation of reference identity. In this work, we introduce MagicAnimate, a diffusion-based framework that aims at enhancing temporal consistency, preserving reference image faithfully, and improving animation fidelity. To achieve this, we first develop a video diffusion model to encode temporal information. Second, to maintain the appearance coherence across frames, we introduce a novel appearance encoder to retain the intricate details of the reference image. Leveraging these two innovations, we further employ a simple video fusion technique to encourage smooth transitions for long video animation. Empirical results demonstrate the superiority of our method over baseline approaches on two benchmarks. Notably, our approach outperforms the strongest baseline by over 38% in terms of video fidelity on the challenging TikTok dancing dataset. Code and model will be made available.

AK

810,649 views • 2 years ago

🔥 Kemi Badenoch has said it clearly — and finally someone is willing to say it outright. Integration is not an administrative box-ticking exercise, and pretending it is has failed Britain. A Conservative government will end identity politics in the state. No hiring. No promotion. No admissions based on protected characteristics. Merit over division. Unity over state-sponsored separation. 🇬🇧

🔥 Kemi Badenoch has said it clearly — and finally someone is willing to say it outright. Integration is not an administrative box-ticking exercise, and pretending it is has failed Britain. A Conservative government will end identity politics in the state. No hiring. No promotion. No admissions based on protected characteristics. Merit over division. Unity over state-sponsored separation. 🇬🇧

Sly U

26,281 views • 5 months ago

imagine if you could just fly like a bird. no missions. no scores. no stress. introducing ☕️🦜 a relaxing, browser-based bird sim where you steer a flock of seagulls in v-formation over the ocean through day, night, storms, rain and lightning. • runs fully in the browser • physics-based flight • smooth on mobile it’s not really a “game”. It’s a quiet and soothing experience. check it out now: ps: this is largely inspired by the flight sim that @levelsio built

imagine if you could just fly like a bird. no missions. no scores. no stress. introducing ☕️🦜 a relaxing, browser-based bird sim where you steer a flock of seagulls in v-formation over the ocean through day, night, storms, rain and lightning. • runs fully in the browser • physics-based flight • smooth on mobile it’s not really a “game”. It’s a quiet and soothing experience. check it out now: ps: this is largely inspired by the flight sim that @levelsio built

kanav

259,504 views • 5 months ago

APEX LEGENDS SEASON 26 🏆OFFICIAL TOP 5 CHANGES 1⃣Caustic Buff: -Gas now removes Aim Assist -Gas now instantly slows enemies -Gas Trap Range increased -Gas removes Ash Dash/Sparrow's Jump -Blue Perks: Gain 2 Traps or Regen HP in gas. -Purple Perks: Tac Charges 300% faster while your standing inside gas. 2⃣Bangalore Rework -Smoke can now be destroyed with grenades -Blue Perk: Smoke heals teammates health -Purple Perk: Fast Revives & Speed boost when running towards knocked teammate. 3⃣Amp Power Ups New inventory "Amp Slot". Any legend can pick up Amp Power Ups from floor loot. -Infinite Ammo -Infinite Batteries: As long as you have 2 batteries in your inventory then you have infinite uses. -Over Armor: Gain 25 extra shield capacity -Power Boost: Reduce Tactical & Ultimate cooldown & restores Tac on kncoked enemy -Overflow Healer: Smalls heals/shields fill entire bar over time. 4⃣Ranked POI Dropships 5⃣RE-45: buffed & now a Burst Energy Pistol 6⃣Ranked Changes - No Heat Shields - Placement > Kills. - Better Skill Based Matchmaking

APEX LEGENDS SEASON 26 🏆OFFICIAL TOP 5 CHANGES 1⃣Caustic Buff: -Gas now removes Aim Assist -Gas now instantly slows enemies -Gas Trap Range increased -Gas removes Ash Dash/Sparrow's Jump -Blue Perks: Gain 2 Traps or Regen HP in gas. -Purple Perks: Tac Charges 300% faster while your standing inside gas. 2⃣Bangalore Rework -Smoke can now be destroyed with grenades -Blue Perk: Smoke heals teammates health -Purple Perk: Fast Revives & Speed boost when running towards knocked teammate. 3⃣Amp Power Ups New inventory "Amp Slot". Any legend can pick up Amp Power Ups from floor loot. -Infinite Ammo -Infinite Batteries: As long as you have 2 batteries in your inventory then you have infinite uses. -Over Armor: Gain 25 extra shield capacity -Power Boost: Reduce Tactical & Ultimate cooldown & restores Tac on kncoked enemy -Overflow Healer: Smalls heals/shields fill entire bar over time. 4⃣Ranked POI Dropships 5⃣RE-45: buffed & now a Burst Energy Pistol 6⃣Ranked Changes - No Heat Shields - Placement > Kills. - Better Skill Based Matchmaking

HisandHersLive

60,913 views • 1 year ago

Most product screenshots are boring. Flat. Static. Forgettable. Screenshot Studio is officially live and it changes that. Formerly @stageeart. Completely rebuilt from the ground up. What started as a simple screenshot editor is now a browser-based studio for screenshots, animated visuals, and export-ready videos. No signup. No watermarks. No artificial limits. New in 2.0: - A faster core and rebuilt UI. - A real timeline editor with tracks + playhead. - 20+ animation presets with keyframes and easing. - MP4, WebM, GIF export. - Hardware-accelerated encoding with WebCodecs (FFmpeg fallback). - High-res exports up to 5x. - Unlimited undo/redo. Drag. Drop. Animate. Export. All in the browser. Proud to share it’s been selected for the Vercel OSS Program. Huge thanks to the Vercel team and Guillermo Rauch for supporting the project. If you’re shipping products and still posting static screenshots, you’re leaving attention on the table. Stage is gone. Screenshot Studio is here. try now link in the comments.

Most product screenshots are boring. Flat. Static. Forgettable. Screenshot Studio is officially live and it changes that. Formerly @stageeart. Completely rebuilt from the ground up. What started as a simple screenshot editor is now a browser-based studio for screenshots, animated visuals, and export-ready videos. No signup. No watermarks. No artificial limits. New in 2.0: - A faster core and rebuilt UI. - A real timeline editor with tracks + playhead. - 20+ animation presets with keyframes and easing. - MP4, WebM, GIF export. - Hardware-accelerated encoding with WebCodecs (FFmpeg fallback). - High-res exports up to 5x. - Unlimited undo/redo. Drag. Drop. Animate. Export. All in the browser. Proud to share it’s been selected for the Vercel OSS Program. Huge thanks to the Vercel team and Guillermo Rauch for supporting the project. If you’re shipping products and still posting static screenshots, you’re leaving attention on the table. Stage is gone. Screenshot Studio is here. try now link in the comments.

Kartik

33,775 views • 6 months ago

We gave AI autonomy over its own body. Left side: a standard wave animation. Every user sees the same thing. Right side: the AI decides how to wave. It controls each limb independently. If the companion is happy? The arc is bigger, slightly faster. Sad? The wave droops. Shy? Smaller, slower. Same gesture. Infinite expressions. The companion's mood, your conversation, the context of the moment, all shape how it physically moves in real time. Now consider this across every gesture, every behavior, every movement. This, all running on a system we built from scratch. At Kindred Labs we're crafting true digital life.

We gave AI autonomy over its own body. Left side: a standard wave animation. Every user sees the same thing. Right side: the AI decides how to wave. It controls each limb independently. If the companion is happy? The arc is bigger, slightly faster. Sad? The wave droops. Shy? Smaller, slower. Same gesture. Infinite expressions. The companion's mood, your conversation, the context of the moment, all shape how it physically moves in real time. Now consider this across every gesture, every behavior, every movement. This, all running on a system we built from scratch. At Kindred Labs we're crafting true digital life.

Max Giammario ∞ KIN

60,623 views • 5 months ago

🚨😳 GAME-CHANGING RED PILL FROM RIPPLE CEO BRAD GARLINGHOUSE’S INTERVIEW “The government owns your identity and Blockchain-based identity is the ONLY way to stop governments from controlling our personal documents.” And XRPL just launched this massive breakthrough. Enter XRPL confidential MPTs + DNA Protocol’s zk-privacy stack: Your sensitive data generates irreversible zero-knowledge proofs anchored directly on the XRPL ledger with $XDNA. ✅ Institutions get private environment, the exact feature crypto industry lacked until now ✅ Verify credentials, age, credentials, or access rights ✅ Zero data exposure ever ✅ Quantum-resistant (no reusable secrets) ✅ Fully private zk-transactions + sovereign identity DNA Protocol executed the FIRST-EVER ZK-Proof Privacy Transaction LIVE on XRPL No more centralized databases. No more surveillance state. No more “your papers please.” This is the infrastructure for true digital freedom: private, permissionless, unstoppable. You now have Private Transactions + Private Identity Verification on XRPL. Epstein mails exposed that this was the reason they saw Ripple as a threat to their ecosystem and Epstein wanted physical control over the keys of ZCash so that they can market it in a disguise of “privacy token.” The sovereign era of privacy is here🛡️🐦‍🔥

🚨😳 GAME-CHANGING RED PILL FROM RIPPLE CEO BRAD GARLINGHOUSE’S INTERVIEW “The government owns your identity and Blockchain-based identity is the ONLY way to stop governments from controlling our personal documents.” And XRPL just launched this massive breakthrough. Enter XRPL confidential MPTs + DNA Protocol’s zk-privacy stack: Your sensitive data generates irreversible zero-knowledge proofs anchored directly on the XRPL ledger with $XDNA. ✅ Institutions get private environment, the exact feature crypto industry lacked until now ✅ Verify credentials, age, credentials, or access rights ✅ Zero data exposure ever ✅ Quantum-resistant (no reusable secrets) ✅ Fully private zk-transactions + sovereign identity DNA Protocol executed the FIRST-EVER ZK-Proof Privacy Transaction LIVE on XRPL No more centralized databases. No more surveillance state. No more “your papers please.” This is the infrastructure for true digital freedom: private, permissionless, unstoppable. You now have Private Transactions + Private Identity Verification on XRPL. Epstein mails exposed that this was the reason they saw Ripple as a threat to their ecosystem and Epstein wanted physical control over the keys of ZCash so that they can market it in a disguise of “privacy token.” The sovereign era of privacy is here🛡️🐦‍🔥

Stellar Rippler🚀

18,489 views • 4 months ago

CSS Tip! 🎠 You can create a responsive infinite marquee animation with container queries and no duplicate items 🤙 li{ animation: slide; } @keyframes slide { to { translate: 0% calc(var(--i) * -100%);}} The trick is animating the items, not the list 😎 More tricks 👇 To get this one working, you need to animate the items and not the list (Watch the video first?). Each item needs to know its row index (--i) in the list and the parent needs to know how many rows are in the list: ul { --count: 12; } li:nth-of-type(1), li:nth-of-type(2) { --i: 0; } li:nth-of-type(3), li:nth-of-type(4) { --i: 1; } Once you have that, translate each item based on its row index in the list li { translate: 0% calc((var(--count) - var(--i)) * 100%); } Now for the animation. The key here is that each row has an animation-delay calculated from its index (--i). That number is offset to make it negative so the animation start is offset ✨ ul { --duration: 10s; } li { --delay: calc((var(--duration) / var(--count)) * (var(--i) - 8)); animation: slide var(--duration) var(--delay) infinite linear; } Make sure to wrap that animation in: @media (prefers-reduced-motion: no-preference) { ... } Lastly, the fun parts! 🤓 To create the "vignette" mask. Use a layered mask on the container 😷 .scene { --buff: 3rem; height: 100%; width: 100%; mask: linear-gradient(transparent, white var(--buff) calc(100% - var(--buff)), transparent), linear-gradient(90deg, transparent, white var(--buff) calc(100% - var(--buff)), transparent); mask-composite: intersect; } To create the 3D skewed effect, use a chained transform (Try toggling it in the demo ⚡️): .grid { transform: rotateX(20deg) rotateZ(-20deg) skewX(20deg); } As for the responsive part, use container queries! 🔥 article { container-type: inline-size; } When the article (card) is narrower than 400px update the grid and animation settings 🤙 Double the rows means double the duration! @container (width < 400px) { .grid { --count: 12; grid-template-columns: 1fr; } li:nth-of-type(1) { --i: 0; } li:nth-of-type(2) { --i: 1; } li:nth-of-type(3) { --i: 2; } li:nth-of-type(4) { --i: 3; } li { --duration: 20s; } } CSS has the magic to be able to update those animations at runtime based on your custom property values 😎 An added bonus in this demo is that it doesn't require any JavaScript at all, for any of it 🤯 We can use CSS :has() for those toggles that update the styles, even the theme toggle! 🫶 Any questions, let me know! Make sure to check out the video. Will do a walkthrough one to follow-up 🤙 CodePen.IO link below! 👇

CSS Tip! 🎠 You can create a responsive infinite marquee animation with container queries and no duplicate items 🤙 li{ animation: slide; } @keyframes slide { to { translate: 0% calc(var(--i) * -100%);}} The trick is animating the items, not the list 😎 More tricks 👇 To get this one working, you need to animate the items and not the list (Watch the video first?). Each item needs to know its row index (--i) in the list and the parent needs to know how many rows are in the list: ul { --count: 12; } li:nth-of-type(1), li:nth-of-type(2) { --i: 0; } li:nth-of-type(3), li:nth-of-type(4) { --i: 1; } Once you have that, translate each item based on its row index in the list li { translate: 0% calc((var(--count) - var(--i)) * 100%); } Now for the animation. The key here is that each row has an animation-delay calculated from its index (--i). That number is offset to make it negative so the animation start is offset ✨ ul { --duration: 10s; } li { --delay: calc((var(--duration) / var(--count)) * (var(--i) - 8)); animation: slide var(--duration) var(--delay) infinite linear; } Make sure to wrap that animation in: @media (prefers-reduced-motion: no-preference) { ... } Lastly, the fun parts! 🤓 To create the "vignette" mask. Use a layered mask on the container 😷 .scene { --buff: 3rem; height: 100%; width: 100%; mask: linear-gradient(transparent, white var(--buff) calc(100% - var(--buff)), transparent), linear-gradient(90deg, transparent, white var(--buff) calc(100% - var(--buff)), transparent); mask-composite: intersect; } To create the 3D skewed effect, use a chained transform (Try toggling it in the demo ⚡️): .grid { transform: rotateX(20deg) rotateZ(-20deg) skewX(20deg); } As for the responsive part, use container queries! 🔥 article { container-type: inline-size; } When the article (card) is narrower than 400px update the grid and animation settings 🤙 Double the rows means double the duration! @container (width < 400px) { .grid { --count: 12; grid-template-columns: 1fr; } li:nth-of-type(1) { --i: 0; } li:nth-of-type(2) { --i: 1; } li:nth-of-type(3) { --i: 2; } li:nth-of-type(4) { --i: 3; } li { --duration: 20s; } } CSS has the magic to be able to update those animations at runtime based on your custom property values 😎 An added bonus in this demo is that it doesn't require any JavaScript at all, for any of it 🤯 We can use CSS :has() for those toggles that update the styles, even the theme toggle! 🫶 Any questions, let me know! Make sure to check out the video. Will do a walkthrough one to follow-up 🤙 CodePen.IO link below! 👇

jhey ʕ•ᴥ•ʔ

542,467 views • 2 years ago

Here's a demo on a project I've been developing and working on for the past 9 months. Called NightBeacon. Using it now in production, getting released fully this week. Our own internally trained models on our own infrastructure (no third party). Trained on our analysts knowledge and behavior (TP/FPs retrain model to be smarter with context). Handles emails (including tonality), attachments, various malicious filetypes (DLL/exe/svg/lnk/etc). Can send it full evtx exports, packet dumps, zip files, whatever. Universal log handler can parse any log from any source, EDR, SIEM, etc. Deep-Scan / sandbox detonation + shellcode emulation with IOC extraction automatically. Automatic playbook generation, full AI-based recommendations custom to the attack. Synthetic training data layer - meaning when it trains on a specific attack at a customer, generates training data based on the customers data but never has any of the actual data or information about the customer in it. No customer information. For areas its weak at, bubbles up and automatically kicks off research to become smarter on a specific topic. Supports GenAI based rulesets (to improve confidence), over 900+ YARA rules, full MITRE ATT&CK integration. Integrated into our SOAR - enriches data, creates playbooks for analysts, MTTR reduces substantially, false positives reduced, true positive escalations. Not using our MDR service? Can integrate into your EDR or SIEM for automatic enrichment and escalation of attacks. Built to help respond faster. More accurately. Be intelligent based on our analysts intelligence. Stop attackers much much faster. Coming soon.. #BinaryDefense

Here's a demo on a project I've been developing and working on for the past 9 months. Called NightBeacon. Using it now in production, getting released fully this week. Our own internally trained models on our own infrastructure (no third party). Trained on our analysts knowledge and behavior (TP/FPs retrain model to be smarter with context). Handles emails (including tonality), attachments, various malicious filetypes (DLL/exe/svg/lnk/etc). Can send it full evtx exports, packet dumps, zip files, whatever. Universal log handler can parse any log from any source, EDR, SIEM, etc. Deep-Scan / sandbox detonation + shellcode emulation with IOC extraction automatically. Automatic playbook generation, full AI-based recommendations custom to the attack. Synthetic training data layer - meaning when it trains on a specific attack at a customer, generates training data based on the customers data but never has any of the actual data or information about the customer in it. No customer information. For areas its weak at, bubbles up and automatically kicks off research to become smarter on a specific topic. Supports GenAI based rulesets (to improve confidence), over 900+ YARA rules, full MITRE ATT&CK integration. Integrated into our SOAR - enriches data, creates playbooks for analysts, MTTR reduces substantially, false positives reduced, true positive escalations. Not using our MDR service? Can integrate into your EDR or SIEM for automatic enrichment and escalation of attacks. Built to help respond faster. More accurately. Be intelligent based on our analysts intelligence. Stop attackers much much faster. Coming soon.. #BinaryDefense

Dave Kennedy

12,905 views • 5 months ago

Meet LongCat-Video-Avatar: a robust audio-driven avatar model that pushes the boundaries of long-form video generation. Compared with the previous InfiniteTalk, LongCat-Video-Avatar delivers far better long-sequence stability and realism. New highlights： ⚙ Built on the LongCat-Video architecture, now supporting Audio-Text-to-Video (AT2V), Audio-Text-Image-to-Video (ATI2V), and Video Continuation modes. 🎭 Open-source SOTA Realism: Ranked 1st in overall anthropomorphism scores for both single and multi-subject scenarios in EvalTalker evaluations (492 participants, 3 independent raters per video). ♾ High-quality long videos: Cross-Chunk Latent Stitching prevents pixel degradation and error accumulation over time, ensuring seamless stitching quality. 🔒 Long-term consistency: Reference Skip Attention maintains ID consistency while eliminating rigid copy-paste artifacts. 🪄 Supports multi-person and infinite-length video generation. 🔗Open-sourced Code: Hugging Face: Project: Paper:

Meituan LongCat

28,170 views • 8 months ago

We are excited to share our work “Event-Aided Sharp Radiance Field Reconstruction for Fast-Flying Drones” published in IEEE Transactions on Robotics IEEE Transactions on Robotics (T-RO), which tackles sharp radiance field reconstruction under agile drone motion, where RGB frames are heavily motion-blurred and pose priors become unreliable! 4 years in the making! Code & dataset released! PDF: Code & Dataset: Full Narrated Video: High-speed flight is essential for time- and battery-constrained missions (e.g., inspection, exploration, search & rescue). However, fast motion corrupts visual data with severe motion blur and introduces drift/noise in visual-inertial odometry, making NeRF-based 3D reconstruction particularly brittle. We propose a unified framework that leverages asynchronous #EventCamera streams together with motion-blurred frames to reconstruct high-fidelity radiance fields from agile drone flights. Our key idea is to embed event-image fusion directly into radiance field optimization while jointly refining a shared, continuous-time camera trajectory initialized from event-based VIO. This enables us to recover sharp radiance fields and accurate trajectories without ground-truth supervision during training. We validate our method on synthetic data and on real sequences captured by a drone flying up to 2 m/s. Despite severe blur and noisy pose priors, our method preserves fine scene details and achieves a performance gain of over 50% on real-world data compared to state-of-the-art methods. Kudos to Rong Zou and Marco Cannici! Marco Cannici Reference: Rong Zou*, Marco Cannici*, Davide Scaramuzza Event-Aided Sharp Radiance Field Reconstruction for Fast-Flying Drones IEEE Transactions on Robotics (T-RO), 2026 NCCR Robotics European Research Council (ERC) AUTOASSESS UZH IfI University of Zurich UZH Science Prophesee SynSense UZH Space Hub

We are excited to share our work “Event-Aided Sharp Radiance Field Reconstruction for Fast-Flying Drones” published in IEEE Transactions on Robotics IEEE Transactions on Robotics (T-RO), which tackles sharp radiance field reconstruction under agile drone motion, where RGB frames are heavily motion-blurred and pose priors become unreliable! 4 years in the making! Code & dataset released! PDF: Code & Dataset: Full Narrated Video: High-speed flight is essential for time- and battery-constrained missions (e.g., inspection, exploration, search & rescue). However, fast motion corrupts visual data with severe motion blur and introduces drift/noise in visual-inertial odometry, making NeRF-based 3D reconstruction particularly brittle. We propose a unified framework that leverages asynchronous #EventCamera streams together with motion-blurred frames to reconstruct high-fidelity radiance fields from agile drone flights. Our key idea is to embed event-image fusion directly into radiance field optimization while jointly refining a shared, continuous-time camera trajectory initialized from event-based VIO. This enables us to recover sharp radiance fields and accurate trajectories without ground-truth supervision during training. We validate our method on synthetic data and on real sequences captured by a drone flying up to 2 m/s. Despite severe blur and noisy pose priors, our method preserves fine scene details and achieves a performance gain of over 50% on real-world data compared to state-of-the-art methods. Kudos to Rong Zou and Marco Cannici! Marco Cannici Reference: Rong Zou, Marco Cannici, Davide Scaramuzza Event-Aided Sharp Radiance Field Reconstruction for Fast-Flying Drones IEEE Transactions on Robotics (T-RO), 2026 NCCR Robotics European Research Council (ERC) AUTOASSESS UZH IfI University of Zurich UZH Science Prophesee SynSense UZH Space Hub

Davide Scaramuzza

12,028 views • 5 months ago

This PV was made by a 60-year-old man, and he's more cyberpunk than most of us will ever be. Meet Uncle Zhao. 30 years of hand animating, now finishing a sci-fi short set in a recycling station town where humans have "evolved" into part-swapping Benchmark Humans just to survive. The full-length version is being redone entirely with Seedance 2.5. he already had over 10 minutes of finished footage when the new model dropped, and scrapped it to redo everything for better quality instead of shipping what he had. It's based on Legend of the Cyber Heroes, a novel on Qidian, and fans have been comparing every frame to the book. Age is just a number when the work looks like this. Full Seedance 2.5 version isn't out yet; comment if you want it sooner #Cyberpunk #animation

This PV was made by a 60-year-old man, and he's more cyberpunk than most of us will ever be. Meet Uncle Zhao. 30 years of hand animating, now finishing a sci-fi short set in a recycling station town where humans have "evolved" into part-swapping Benchmark Humans just to survive. The full-length version is being redone entirely with Seedance 2.5. he already had over 10 minutes of finished footage when the new model dropped, and scrapped it to redo everything for better quality instead of shipping what he had. It's based on Legend of the Cyber Heroes, a novel on Qidian, and fans have been comparing every frame to the book. Age is just a number when the work looks like this. Full Seedance 2.5 version isn't out yet; comment if you want it sooner #Cyberpunk #animation

Parul Gautam

101,205 views • 23 days ago

One prompt. Full kids cartoon movie. Over $10,000 per month. This is actually ******* insane... No face. No real animation skills. No team. Just 3 stupidly simple steps: 1. Pick a high-CPM niche (kids stories) 2. Go to invidio io v3.0 3. Type a prompt like “Titanic but make it kid-friendly Disney Pixar style” - and it generates a full-length animated video ready to upload. One prompt = complete cartoon movie that kids watch on repeat for hours. Highest CPM niche on the platform. Parents leave it playing all day. While you're still out here grinding charts and hoping for a miracle 10x... This is the laziest, most unfair advantage in content creation right now. The tool is so good it feels like cheating. If this one actually hit you, smash like and follow me. I’m exposing every raw AI money printer and faceless system that’s working stupidly well in 2026.

One prompt. Full kids cartoon movie. Over $10,000 per month. This is actually ******* insane... No face. No real animation skills. No team. Just 3 stupidly simple steps: 1. Pick a high-CPM niche (kids stories) 2. Go to invidio io v3.0 3. Type a prompt like “Titanic but make it kid-friendly Disney Pixar style” - and it generates a full-length animated video ready to upload. One prompt = complete cartoon movie that kids watch on repeat for hours. Highest CPM niche on the platform. Parents leave it playing all day. While you're still out here grinding charts and hoping for a miracle 10x... This is the laziest, most unfair advantage in content creation right now. The tool is so good it feels like cheating. If this one actually hit you, smash like and follow me. I’m exposing every raw AI money printer and faceless system that’s working stupidly well in 2026.

Frogify

68,213 views • 2 months ago

Future CSS Tip! 🍏 You can create this Apple-style photo scroller by combining CSS scroll-driven animations and CSS scroll-snap 😍 Peep those changing captions 👀 No JS! img { animation: highlight both linear; animation-timeline: view(inline); 👈 Horizontal animation-range: cover 0% cover 50%; 👈 Finish } @ keyframes highlight { 50% { translate: 0 0; scale: var(--starting-scale); 👈 props opacity: var(--starting-opacity); 👈 } 100% { translate: 0 0; scale: 1; opacity: 1; } } Without the animation support, you get a standard unordered list containing some s 🤙 How do we swap the captions though? The "trick" is to use position: absolute on the figcaption and animate their opacity based on the ViewTimeline of their parent list item 😎 figcaption { animation: show both linear; animation-timeline: --list-item; } @ keyframes show { 0%, 45%, 55%, 100% { opacity: 0; } 50% { opacity: 1; } } li { view-timeline-name: --list-item; view-timeline-axis: inline; 👈 important! } The parent of the scroll track uses position: relative so all the captions sit in the middle even though they are in the right place for the markup 🙌 The last bit is the scroll-snap 🤙 Not much to it at all. Wrap the list and make it scrollable. Then add scroll-snap-type .wrapper { scroll-snap-type: x mandatory; overflow-x: scroll; } Then make sure each list item has scroll-snap-align set on it li { scroll-snap-align: center; } That's it! Pretty cool demo to put together and see how to do this stuff with these APIs 🤓 A lot of cool little tricks to pick up for writing your CSS! ⭐️ CodePen.IO link below! 👇

Future CSS Tip! 🍏 You can create this Apple-style photo scroller by combining CSS scroll-driven animations and CSS scroll-snap 😍 Peep those changing captions 👀 No JS! img { animation: highlight both linear; animation-timeline: view(inline); 👈 Horizontal animation-range: cover 0% cover 50%; 👈 Finish } @ keyframes highlight { 50% { translate: 0 0; scale: var(--starting-scale); 👈 props opacity: var(--starting-opacity); 👈 } 100% { translate: 0 0; scale: 1; opacity: 1; } } Without the animation support, you get a standard unordered list containing some s 🤙 How do we swap the captions though? The "trick" is to use position: absolute on the figcaption and animate their opacity based on the ViewTimeline of their parent list item 😎 figcaption { animation: show both linear; animation-timeline: --list-item; } @ keyframes show { 0%, 45%, 55%, 100% { opacity: 0; } 50% { opacity: 1; } } li { view-timeline-name: --list-item; view-timeline-axis: inline; 👈 important! } The parent of the scroll track uses position: relative so all the captions sit in the middle even though they are in the right place for the markup 🙌 The last bit is the scroll-snap 🤙 Not much to it at all. Wrap the list and make it scrollable. Then add scroll-snap-type .wrapper { scroll-snap-type: x mandatory; overflow-x: scroll; } Then make sure each list item has scroll-snap-align set on it li { scroll-snap-align: center; } That's it! Pretty cool demo to put together and see how to do this stuff with these APIs 🤓 A lot of cool little tricks to pick up for writing your CSS! ⭐️ CodePen.IO link below! 👇

jhey ʕ•ᴥ•ʔ

234,561 views • 2 years ago

JUST NOW: U.S. Secretary of WAR Pete Hegseth delivered a blunt message at the Shangri-La Dialogue: the era of America carrying its allies is over. 🇺🇸👏 Secretary of WAR made it clear that alliances are measured by military capability, not symbolism, stressing the need for more ships, submarines, and combat readiness over diplomatic rhetoric. 💪 CHINA WARNING: Reaffirming America as a Pacific power, Hegseth declared, "Make no mistake: America is a Pacific nation," and insisted that China respect longstanding U.S. positions in the region. 🇨🇳 PRIORITY FOR "MODEL ALLIES": Pete made it clear that countries that invest in their own defense will receive faster access to U.S. weapons and deeper industrial cooperation. EUROPE & NATO ON NOTICE: Hegseth criticized European nations for underinvesting in defense while pursuing open-border policies. As Pete said, "For too long, polite pleas from our European allies to spend more on their own defense fell on deaf ears." 🇪🇺 The message was clear future alliances will be built on military strength, burden-sharing, and measurable commitments—not promises alone. 🦅 (minimal AI visuals (e.g., flags and animations); rest fact-based coverage)

JUST NOW: U.S. Secretary of WAR Pete Hegseth delivered a blunt message at the Shangri-La Dialogue: the era of America carrying its allies is over. 🇺🇸👏 Secretary of WAR made it clear that alliances are measured by military capability, not symbolism, stressing the need for more ships, submarines, and combat readiness over diplomatic rhetoric. 💪 CHINA WARNING: Reaffirming America as a Pacific power, Hegseth declared, "Make no mistake: America is a Pacific nation," and insisted that China respect longstanding U.S. positions in the region. 🇨🇳 PRIORITY FOR "MODEL ALLIES": Pete made it clear that countries that invest in their own defense will receive faster access to U.S. weapons and deeper industrial cooperation. EUROPE & NATO ON NOTICE: Hegseth criticized European nations for underinvesting in defense while pursuing open-border policies. As Pete said, "For too long, polite pleas from our European allies to spend more on their own defense fell on deaf ears." 🇪🇺 The message was clear future alliances will be built on military strength, burden-sharing, and measurable commitments—not promises alone. 🦅 (minimal AI visuals (e.g., flags and animations); rest fact-based coverage)

Donald J Trump Posts TruthSocial

114,861 views • 2 months ago