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.

We put a neural network inside the camera pixel. Introducing RETINA: a vision sensor that runs AI where the light lands

Synapse Semiconductor

1,202 subscribers

207,914 views • 1 month ago •via X (Twitter)

Arts Science & Technology

Anya Rossi• Live Now

Private livecam show

35 Comments

Sanjeev Chauhan1 month ago

We've designed some neat camera pixels that can run a CNN layer (matrix-matrix multiplication) using trapped charges in the pixel itself. These act as an image filter, letting only the important information in.

Cole1 month ago

I’m interested in building mini robots the size of baseballs to perform basic physical assistance tasks (camera / rear visibility, projection, and general fun at low cost). Lmk if yall wanna talk

Aidan Tiruvan1 month ago

better vision sensors def needed for the future of robotics. congrats!

Nolan Connolly1 month ago

Love the Steve Jobs iPhone reference

3DTOPO1 month ago

Uhhh Retina is a registered trademark by Apple

Roger Douglas Ware1 month ago

Imagine ai as a nural linguistic communication standard for pre cognitive lacing. Conciousness on a theta gamma band ai on an alpha/ beta channel. I'm putting ai into the stream of consciousness. Unless someone comes back in time to say stop it. Gpl-2 my work, my heart is rich. Money is discretionary. Play well with others.

Edward1 month ago

congrats!

devesh1 month ago

Holy banger 🔥

William Chen1 month ago

Life changing.

Daniel Keller1 month ago

Diffractive optical neural net?

Jeremy1 month ago

Can u put a chip in my eyes

Raul Izahi Lopez 🇺🇸1 month ago

Ready to make tens of millions of them?

Niki Kotecha1 month ago

love this guys!!!

steve1 month ago

Woa now

Kartikey Pandey1 month ago

@b1shtream Nice

Con1 month ago

Hate to say it but, I’ll believe it when I see it!!

Sarah Zhao1 month ago

LAUNCH DAY

Victor1 month ago

Let’s go 🚀

Dhruv Doshi1 month ago

Would love to learn more about this

Bruno Lima1 month ago

Let’s go!!

Ricky Chan1 month ago

Synapse, nice name.

Aubteen Pour-Biazar1 month ago

👏👏

Alex Pospekhov1 month ago

Insanely sick !!!!!!!!!!!!!!!!! Woah!!!!!

use.fo - AI Voice Typing1 month ago

The key is not just running AI on the sensor, but turning those raw signals into structured, usable language in real-time.

Kinjal Goswami1 month ago

Edge AI keeps getting crazier.

Adrian Thönig1 month ago

😂😂😂😂😂😂😂 bro that's exactly the steve jobs marketing pattern from the first iphone release.

LoBa._1 month ago

Coolest thing I’ve seen today

Priyav K Kaneria1 month ago

waow

Rey Neill1 month ago

Sorry?

Shubham1 month ago

When would be the best time to place an order? I'm looking forward to get one for a project I'm working on.

James Sackl1 month ago

Cool!

Tanay Lohia1 month ago

That be music is too distracting

homuraakemifan1 month ago

Stacking a camera sensor on top of an AI chip? Isn't that what China already does? Is it really that big of a deal that America is knocking it off? I guess the AI ballistic missile targeting system is very strong 🤔

Vihaan Shringi1 month ago

Cool

NovaBishop1 month ago

😳

Related Videos

A peek inside the mind of AI through neural network visualization.

A peek inside the mind of AI through neural network visualization.

Massimo

60,161 views • 7 months ago

Best visual I’ve seen of the inside of a Neural network, a CNN to be precise.

Best visual I’ve seen of the inside of a Neural network, a CNN to be precise.

Linus ✦ Ekenstam

174,542 views • 1 year ago

Introducing the next Raspberry Pi. An Ai model that runs on a credit card.

Introducing the next Raspberry Pi. An Ai model that runs on a credit card.

Founders Inc

157,302 views • 1 year ago

A Neural Network Can Grow New Neurons Where It Is Confused? In 1994, Bernd Fritzke published A Growing Neural Gas Network Learns Topologies. He introduced a network that starts small, follows incoming data, and inserts new neurons where its error is highest. In the animation, the fog is the drifting data. The glowing nodes are neurons. The fibers are learned connections. The network grows into a living skeleton of the manifold.

A Neural Network Can Grow New Neurons Where It Is Confused? In 1994, Bernd Fritzke published A Growing Neural Gas Network Learns Topologies. He introduced a network that starts small, follows incoming data, and inserts new neurons where its error is highest. In the animation, the fog is the drifting data. The glowing nodes are neurons. The fibers are learned connections. The network grows into a living skeleton of the manifold.

Mathelirium

38,431 views • 4 months ago

How far can a very simple eye go in solving vision tasks? Like a 1-pixel camera? Humans have one of the greatest eyes in nature, while many animals have significantly simpler eyes and visual systems yet show complex perceptual behavior. In an interesting project, we find that many computer vision tasks can be solved without a typical camera and with such simple 1-pixel sensors (photoreceptors). We also find that proper design (e.g., where to place the photoreceptors strategically) makes a big difference, so we developed a computational design method to find them. 🌐 👁️[Solving Vision Tasks with Simple Photoreceptors Instead of Cameras] 🧵1/n

How far can a very simple eye go in solving vision tasks? Like a 1-pixel camera? Humans have one of the greatest eyes in nature, while many animals have significantly simpler eyes and visual systems yet show complex perceptual behavior. In an interesting project, we find that many computer vision tasks can be solved without a typical camera and with such simple 1-pixel sensors (photoreceptors). We also find that proper design (e.g., where to place the photoreceptors strategically) makes a big difference, so we developed a computational design method to find them. 🌐 👁️[Solving Vision Tasks with Simple Photoreceptors Instead of Cameras] 🧵1/n

Amir Zamir

75,908 views • 2 years ago

OAK 4 CS now supports PTP synchronization. Sync multi-camera exposures over standard Ethernet. No FSYNC wiring, no daisy chains, no trigger cables across long distances. Just a shared clock distributed over the network. For multi-camera perception, distributed AI vision, and sensor fusion at scale.

OAK 4 CS now supports PTP synchronization. Sync multi-camera exposures over standard Ethernet. No FSYNC wiring, no daisy chains, no trigger cables across long distances. Just a shared clock distributed over the network. For multi-camera perception, distributed AI vision, and sensor fusion at scale.

Luxonis | Robotic Vision

32,787 views • 3 months ago

A single exchange––one that unifies liquidity, connects every network, and powers applications everywhere. One that can infinitely scale horizontally & vertically. With MetaDEX03 we finally have the OS to make that vision a reality: Introducing Aero. Coming soon to Ethereum.

A single exchange––one that unifies liquidity, connects every network, and powers applications everywhere. One that can infinitely scale horizontally & vertically. With MetaDEX03 we finally have the OS to make that vision a reality: Introducing Aero. Coming soon to Ethereum.

Dromos

417,732 views • 10 months ago

Introducing Manako, the fastest way to turn any camera into an vision ai agent. Go on Join our waitlist.

Introducing Manako, the fastest way to turn any camera into an vision ai agent. Go on Join our waitlist.

Manako

26,499 views • 3 months ago

AI FOOTBALL ANALYSIS. A FULL COMPUTER VISION SYSTEM. BUILT ON YOLO, OPENCV, AND PYTHON. You upload a regular match video. No sensors, no GPS trackers, just camera footage. The neural network finds every player, referee, and ball on its own. Every frame, in real time. KMeans clustering breaks down jersey colors pixel by pixel. The system splits players into teams automatically. Without a single manual hint. Optical Flow tracks camera movement. Separates it from player movement. Perspective Transformation converts pixels into real meters. Speed of every player. Distance covered. Ball possession percentage. All calculated automatically. Four hours of tutorial from zero to a working system. The model is trained on real Bundesliga matches. Runs on a regular GPU. Python code - take it and run. Sports analytics is no longer behind closed doors. AI leveled the playing field.

AI FOOTBALL ANALYSIS. A FULL COMPUTER VISION SYSTEM. BUILT ON YOLO, OPENCV, AND PYTHON. You upload a regular match video. No sensors, no GPS trackers, just camera footage. The neural network finds every player, referee, and ball on its own. Every frame, in real time. KMeans clustering breaks down jersey colors pixel by pixel. The system splits players into teams automatically. Without a single manual hint. Optical Flow tracks camera movement. Separates it from player movement. Perspective Transformation converts pixels into real meters. Speed of every player. Distance covered. Ball possession percentage. All calculated automatically. Four hours of tutorial from zero to a working system. The model is trained on real Bundesliga matches. Runs on a regular GPU. Python code - take it and run. Sports analytics is no longer behind closed doors. AI leveled the playing field.

zostaff

1,076,853 views • 5 months ago

Introducing the new Pixel 10a. It has everything you need — like a completely flat, durable design, more battery life, advanced AI tools and a best-in-class camera system* — for just $499. *Among smartphones under $500 (MSRP) in the U.S. Based on third-party evaluation of Pixel 10a.

Introducing the new Pixel 10a. It has everything you need — like a completely flat, durable design, more battery life, advanced AI tools and a best-in-class camera system* — for just $499. *Among smartphones under $500 (MSRP) in the U.S. Based on third-party evaluation of Pixel 10a.

Google

84,511 views • 7 months ago

Today, Red Hat completed the acquisition of Red Hat AI (formerly Neural Magic), a pioneer in software and algorithms that accelerate #GenAI inference workloads. Read how we are accelerating our vision for #AI’s future:

Today, Red Hat completed the acquisition of Red Hat AI (formerly Neural Magic), a pioneer in software and algorithms that accelerate #GenAI inference workloads. Read how we are accelerating our vision for #AI’s future:

Red Hat

29,613 views • 1 year ago

Introducing Neural Tensor Dynamics (NTD): $NTD integrates a unique staking solution with high-yield $wTAO rewards, AI-driven DeFi tools, and validator services, all built atop the Bittensor network. Let’s explore these features in detail:

Introducing Neural Tensor Dynamics (NTD): $NTD integrates a unique staking solution with high-yield $wTAO rewards, AI-driven DeFi tools, and validator services, all built atop the Bittensor network. Let’s explore these features in detail:

Neural Tensor Dynamics

51,995 views • 2 years ago

Can a single neural network policy generalize over poses, objects, obstacles, backgrounds, scene arrangements, in-hand objects, and start/goal states? Introducing Neural MP: A generalist policy for solving motion planning tasks in the real world 🤖 1/N

Can a single neural network policy generalize over poses, objects, obstacles, backgrounds, scene arrangements, in-hand objects, and start/goal states? Introducing Neural MP: A generalist policy for solving motion planning tasks in the real world 🤖 1/N

Murtaza Dalal

114,720 views • 2 years ago

A film that you can put over your camera that will take invisible, infrared light photons, and perform an anti-Stokes effect and emit a visible red light photon wherever infrared light exists. Notice the green space!

A film that you can put over your camera that will take invisible, infrared light photons, and perform an anti-Stokes effect and emit a visible red light photon wherever infrared light exists. Notice the green space!

Roger Seheult, MD

16,274 views • 1 year ago

Implemented a visualization for the neural network today. This is a neural network being trained to learn the snake game in a terminal. It's built in Rust using Ratatui More input details in thread

Implemented a visualization for the neural network today. This is a neural network being trained to learn the snake game in a terminal. It's built in Rust using Ratatui More input details in thread

Bones

165,715 views • 2 years ago

Does GPT understand the world? Here is what Ilya Sutskever, co-founder of OpenAI, says during a discussion with Jensen Huang, CEO of Nvidia: (1) When we train a large neural network to accurately predict the next word in lots of different texts from the internet, the AI is learning a world model. (2) On the surface, it may look like learning correlations in text, but it turns out that to 'just learn' statistical correlations in text, to compress information really well, what the neural network learns is some representation of the process that produced the text. (3) This text is a projection of the world...what the neural network is learning is aspects of the world, of people, of the human conditions, their hopes, dreams, motivations, their interactions...the situations we are in. The neural network learns a compressed, abstract, usable representation." Do you think learning representations = understanding? Are large language models simply stochastic parrots, or are they much more?

Does GPT understand the world? Here is what Ilya Sutskever, co-founder of OpenAI, says during a discussion with Jensen Huang, CEO of Nvidia: (1) When we train a large neural network to accurately predict the next word in lots of different texts from the internet, the AI is learning a world model. (2) On the surface, it may look like learning correlations in text, but it turns out that to 'just learn' statistical correlations in text, to compress information really well, what the neural network learns is some representation of the process that produced the text. (3) This text is a projection of the world...what the neural network is learning is aspects of the world, of people, of the human conditions, their hopes, dreams, motivations, their interactions...the situations we are in. The neural network learns a compressed, abstract, usable representation." Do you think learning representations = understanding? Are large language models simply stochastic parrots, or are they much more?

Alex Ker 🔭

1,367,117 views • 2 years ago