View allAll Photos Tagged ComputerVision
Camera manufacturers are meeting ever-increasing visual world demands seamlessly with the help of Computer Vision models. Data Labeler is the best place for high-quality personalized labeled data sets. Read more - www.datalabeler.com/camera-manufacturers-are-making-the-b...
for more - www.kitdastudio.com/?p=90
i’ve joint the openKinect project initialized by theo, memo (openframeworks forum) at 2010 nov.
this device (xBox kinect sensor) is superb, it got an IR beam generator enabling the depth-camera to read the IR reflection
(so it works perfectly indoor in day time or in complete darkness),
and it got a normal web cam for more video-programming manipulation (like face detection etc)
this project is so exciting, it is another level of computer vision to me.
it just opened up unlimited possibilities in interactive installations !
what to do next ? in order to make it applicable to interactive installation,
it has to be able to do some data things
- limbs motion detection
- incorporate openCV face detection
- dynamic range human-shape recognition
- hand / finger motion recognition
- sync the 2 cams as well as possible
From left, Associate Professor of Electrical Engineering and Computer Science Alanson Sample, and PhD student Yasha Iravanchi examine their prototype PrivacyLens in the Interactive Sensing and Computing Lab at the Bob and Betty Beyster Building, on the North Campus of the University of Michigan in Ann Arbor on Tuesday, July 9, 2024. The ISC is part of the Human Centered Computing Laboratory.
This phototype camera, called PrivacyLens, was designed by Yasha Iravantchi, a doctoral student of computer science and engineering, and Alanson Sample, an associate professor of computer science and engineering, to protect people's privacy in devices that use cameras for sensing. Such devices, including roombas, automated vehicles, and home assistant technologies, such as Alexa, use cameras to avoid collisions or monitor health, fitness, and activity in the home. Those same cameras increase the risk of sensitive, personal information leaking on the internet. PrivacyLens replaces humans with stick figures in the recorded images, so that the devices can register that people are present without putting an individual's privacy at risk.
Photo: Brenda Ahearn/University of Michigan, College of Engineering, Communications and Marketing
Photo: Francesca Centonze
Longevity of light bulbs (and how to make them last longer)
Electroacoustic interactive performance
Stefano D'Alessio
Vienna 2016
Longevity of light bulbs fuses natural and processed sound with physical movement and dynamic light.
The performance proposes a DIY way of music making, focusing on low budget items hacking, creative use of new technologies designing alternative interfaces for sound creation.
Resembling of a solo concert piece, L.O.L.B. has as its core an amplified ikea desk lamp, explored in all its sonorities by one performer playing it with bare hands.
An important characteristic of the lamp is that it can be easily moved, this mobility amplifies the physical performance and modifies the original sound, thanks to a custom made interface and digital audio processing.
The interface between the lamp movement and the sound manipulation, is ideated hypothesising how a “natural” sound would behave, taking advantage of the possibilities of digital audio but bringing them out of the computer, interfacing them with something more physical and “primitive” than a controller fader or a mouse.
The focus is to magnify how the sound processing is controlled, transforming unperceivable digits unseeable to human eyes, into visible performative actions.
The spatial position of the light source is tracked by a camera and translated into parameters usable by the audio engine, naturally connecting the lamp movement to the behaviour of its sound transformations.
The final audio, even when processed, has tones which are close to the natural ones, as the audio processes basically consist in different applications audio delay, meaning that there is no synthesised sound or recorded sample added to the final output.
The only light source present in the performance is the lamp’s light bulb itself. Meaning that the moving lamp is, beside changing its own sound, acting as a dynamic lighting device. It points and looks in different direction, behaving like a curious creature, illuminating different part of the environment, scanning through the audience and sometimes blinding the people.
The result is a performance and electroacoustic piece, where natural and artificial tones fuse seamlessly, symbiotically combined with light, developing in time with movements of the performer and its instrument.
Detail photo of the prototype of the PrivacyLens which places two lenses side-by-side, one that records traditional RGB images and another that records images based on thermal data in the Interactive Sensing and Computing Lab at the Bob and Betty Beyster Building, on the North Campus of the University of Michigan in Ann Arbor on Tuesday, July 9, 2024.
This phototype camera, called PrivacyLens, was designed by Yasha Iravantchi, a doctoral student of computer science and engineering, and Alanson Sample, an associate professor of computer science and engineering, to protect people's privacy in devices that use cameras for sensing. Such devices, including roombas, automated vehicles, and home assistant technologies, such as Alexa, use cameras to avoid collisions or monitor health, fitness, and activity in the home. Those same cameras increase the risk of sensitive, personal information leaking on the internet. PrivacyLens replaces humans with stick figures in the recorded images, so that the devices can register that people are present without putting an individual's privacy at risk.
Photo: Brenda Ahearn/University of Michigan, College of Engineering, Communications and Marketing
From left, Associate Professor of Electrical Engineering and Computer Science Alanson Sample, and PhD student Yasha Iravanchi examine their prototype PrivacyLens in the Interactive Sensing and Computing Lab at the Bob and Betty Beyster Building, on the North Campus of the University of Michigan in Ann Arbor on Tuesday, July 9, 2024. The ISC is part of the Human Centered Computing Laboratory.
This phototype camera, called PrivacyLens, was designed by Yasha Iravantchi, a doctoral student of computer science and engineering, and Alanson Sample, an associate professor of computer science and engineering, to protect people's privacy in devices that use cameras for sensing. Such devices, including roombas, automated vehicles, and home assistant technologies, such as Alexa, use cameras to avoid collisions or monitor health, fitness, and activity in the home. Those same cameras increase the risk of sensitive, personal information leaking on the internet. PrivacyLens replaces humans with stick figures in the recorded images, so that the devices can register that people are present without putting an individual's privacy at risk.
Photo: Brenda Ahearn/University of Michigan, College of Engineering, Communications and Marketing