{"id":176048,"date":"2021-02-11T22:37:00","date_gmt":"2021-02-11T19:37:00","guid":{"rendered":"https:\/\/en.buradabiliyorum.com\/new-ai-technique-lets-robots-detect-human-touch-by-analyzing-shadows\/"},"modified":"2021-02-11T22:37:00","modified_gmt":"2021-02-11T19:37:00","slug":"new-ai-technique-lets-robots-detect-human-touch-by-analyzing-shadows","status":"publish","type":"post","link":"https:\/\/buradabiliyorum.com\/en\/new-ai-technique-lets-robots-detect-human-touch-by-analyzing-shadows\/","title":{"rendered":"#New AI technique lets robots detect human touch by analyzing shadows"},"content":{"rendered":"

“#New AI technique lets robots detect human touch by analyzing shadows<\/strong>”<\/p>\n

\n Scientists from Cornell University have developed a way for\u00a0robots to identify physical interactions\u00a0just by analyzing a user\u2019s shadows.<\/p>\n

Their ShadowSense system\u00a0uses an off-the-shelf USB camera to capture the shadows produced by hand gestures on a robot\u2019s skin.\u00a0Algorithms then classify the movements to infer the user\u2019s specific interaction.<\/p>\n

Study lead author Guy Hoffman said<\/a> the method provides a natural way of interacting with robots without relying on large and costly sensor arrays:<\/p>\n

Touch is such an important mode of communication for most organisms, but it has been virtually absent from human-robot interaction. One of the reasons is that full-body touch used to require a massive number of sensors, and was therefore not practical to implement. This research offers a low-cost alternative.<\/p>\n<\/blockquote>\n

The researchers tried out the system on an inflatable robot with a camera underneath its skin.<\/p>\n

[Read:\u00a0How Polestar is using blockchain to increase transparency]<\/em><\/p>\n

They trained and tested the classification algorithms with shadow images of six gestures:\u00a0touching with a palm, punching, touching with two hands, hugging, pointing, and not touching.<\/p>\n

It successfully distinguished between the gestures with\u00a087<\/span>.5 \u2212 96% <\/span>accuracy, depending <\/span>on the <\/span>lighting.<\/span><\/p>\n

\"The
Credit: Hu et al.<\/figcaption>
<\/a>ShadowSense was most accurate in daylight (96%), followed by dusk (93%), and night (87%).<\/figcaption><\/figure>\n

The researchers envision mobile guide robots using the tech to respond to different gestures, such as turning to face a human when it detects a poke, and moving away when it senses a tap on the back.<\/p>\n

It could also add\u00a0interactive touch screens to inflatable robots and make home assistant droids more privacy-friendly<\/span>.<\/p>\n

\u201cIf the robot can only see you in the form of your shadow, it can detect what you\u2019re doing without taking high fidelity images of your app<\/a>earance,\u201d said Hoffman. \u201cThat gives you a physical filter and protection, and provides psychological comfort.\u201d<\/p>\n

You can read the full study paper here<\/a>.<\/p>\n

\n Published February 11, 2021 \u2014 19:37 UTC<\/p><\/div>\n