Trending...
- California: Governor Newsom proclaims Independence Day 2025
- New Release: 'The Invisible Alternative' Unveiled by Atrisk Corporation, Resilient
- Transgender Visual Artist Debuts New Work
LOS ANGELES - Californer -- A new research paper published in eLight demonstrated a new paradigm to achieve privacy-preserving imaging by building a fundamentally new type of imager designed by AI. In their paper, UCLA researchers presented a smart camera design that images only certain types of desired objects, while instantaneously erasing other types of objects from its images without requiring any digital processing.
This new camera design consists of successive transmissive surfaces, each composed of tens of thousands of diffractive features at the scale of the wavelength of light. The structure of these transmissive surfaces is optimized using deep learning to modulate the phase of the transmitted optical fields such that the camera only images certain types/classes of desired objects and erases the others. After its deep learning-based design (training), the resulting layers are fabricated and assembled in 3D. When the input objects from the target classes of objects appear in front of this camera, they form high-quality images at the camera's output – as desired. In contrast, when the input objects in front of the same camera belong to other undesired classes, they are optically erased, forming non-informative and low-intensity patterns similar to random noise. Since the characteristic information of undesired classes of objects is all-optically erased at the camera output through light diffraction, this AI-designed camera never records their direct images. Therefore, the protection of privacy is maximized since an adversarial attack that has access to the recorded images of this camera cannot bring the information back.
More on The Californer
This AI-based camera design was also used to build encryption cameras, providing an additional layer of security and privacy protection. Such an encryption camera optically performs a selected linear transformation, exclusively for the target objects of interest. Only those with access to the decryption key (i.e., the inverse linear transformation in this case) can recover the original image of the target objects. On the other hand, the information of the other undesired objects is irreversibly lost since the AI-designed camera all-optically erases them at the output. Therefore, even if the decryption key is applied to the recorded images, it yields noise-like, unrecognizable features for other classes of undesired objects.
This research was led by Professor Aydogan Ozcan along with Professor Mona Jarrahi, both from the Electrical and Computer Engineering (ECE) department at UCLA. The other authors include graduate students Bijie Bai, Yi Luo, Tianyi Gan, Yuhang Li, Yifan Zhao, Deniz Mengu and post-doctoral researcher Dr. Jingtian Hu, all with the ECE department at UCLA. The authors report the support of the US Office of Naval Research and the US Department of Energy.
More on The Californer
See the article: https://elight.springeropen.com/articles/10.1186/s43593-022-00021-3
This new camera design consists of successive transmissive surfaces, each composed of tens of thousands of diffractive features at the scale of the wavelength of light. The structure of these transmissive surfaces is optimized using deep learning to modulate the phase of the transmitted optical fields such that the camera only images certain types/classes of desired objects and erases the others. After its deep learning-based design (training), the resulting layers are fabricated and assembled in 3D. When the input objects from the target classes of objects appear in front of this camera, they form high-quality images at the camera's output – as desired. In contrast, when the input objects in front of the same camera belong to other undesired classes, they are optically erased, forming non-informative and low-intensity patterns similar to random noise. Since the characteristic information of undesired classes of objects is all-optically erased at the camera output through light diffraction, this AI-designed camera never records their direct images. Therefore, the protection of privacy is maximized since an adversarial attack that has access to the recorded images of this camera cannot bring the information back.
More on The Californer
- Easton & Easton, LLP Files Suit Against The Dwelling Place Anaheim & Vineyard USA Over Abuse Allegations
- AI Visibility: The Key to Beating Google's AI Overviews and Regaining Traffic
- First Partner highlights apprenticeship program helping underrepresented youth break into careers in California's iconic entertainment industry
- Stuck Doing Math or Figuring Out Life's Numbers? Calculator.now Makes It Stupidly Simple
- Cal State LA secures funding for two artificial intelligence projects from CSU
This AI-based camera design was also used to build encryption cameras, providing an additional layer of security and privacy protection. Such an encryption camera optically performs a selected linear transformation, exclusively for the target objects of interest. Only those with access to the decryption key (i.e., the inverse linear transformation in this case) can recover the original image of the target objects. On the other hand, the information of the other undesired objects is irreversibly lost since the AI-designed camera all-optically erases them at the output. Therefore, even if the decryption key is applied to the recorded images, it yields noise-like, unrecognizable features for other classes of undesired objects.
This research was led by Professor Aydogan Ozcan along with Professor Mona Jarrahi, both from the Electrical and Computer Engineering (ECE) department at UCLA. The other authors include graduate students Bijie Bai, Yi Luo, Tianyi Gan, Yuhang Li, Yifan Zhao, Deniz Mengu and post-doctoral researcher Dr. Jingtian Hu, all with the ECE department at UCLA. The authors report the support of the US Office of Naval Research and the US Department of Energy.
More on The Californer
- Colbert Packaging Announces WBENC Recognition
- New Mobile Car Detailing Platform Connects Drivers with On-Demand Local Pros
- Over the past three months, California seized $476 million worth of unlicensed cannabis products
- California scores more clean energy records: 9 in 10 days this year partially powered by 100% clean energy
- "Mobile Suit Gundam" Takes Over San Diego Comic-Con 2025
See the article: https://elight.springeropen.com/articles/10.1186/s43593-022-00021-3
Source: UCLA ITA
Filed Under: Science
0 Comments
Latest on The Californer
- Voices for Humanity Builds a Better Future for Ghana with Richmond Atta-Williams
- Haunted Doll House at The Gregangelo Museum
- Mystical Insights at the Gregangelo Museum
- Ben Pouladian, CEO of BEP Holdings, to Join ISPE Los Angeles Biomanufacturing Panel
- Raksmart: Promoting Security in Global Hosting Service
- $14M Expansion Deal with Famed David Lloyd Highlights Rebrand of Sports, Entertainment and Gaming Innovation by AI Driven, Online Fan Engagement Co
- Local Pub Grows Unique Community Ownership Program
- Heartfelt Dreams Foundation Launches Campaign to Build CHD Hospital
- California: Acting Governor Eleni Kounalakis proclaims Disability Pride Month
- CHRIS EDMONDS JOINS CITY NATIONAL BANK AS EXECUTIVE VICE PRESIDENT, HEAD OF COMMERCIAL AND NATIONAL CORPORATE BANKING
- Radarsign Tackles Intersection Safety with Launch of Grid-Free Solar LED Stop Sign
- California: $35 million to law enforcement partners to keep communities safe from impaired driving
- Curtis Sergeant's Book The Only One Continues to Equip Christians to Live Fully in by and for God
- IFYA Announces New Froyo Heaven Art Contest
- BKM Capital Partners Publishes Case Study on Hughes Airport Center, Showcasing Strategic Value-Add Execution in Las Vegas
- Miami Real Estate Agent Drastically Increases Interest In Homes
- Adostics & Genmega Announce the Introduction of A-POD
- Auto Careers Online Secures Top Dealers' Choice Award for Eighth Consecutive Year
- MANTL Becomes First Fintech to Offer Plaid Layer
- LIB and Nidec Rejoin Forces for Giant TH-0098 Temperature Humidity Test Chamber