Project:Shitcam: Difference between revisions
From London Hackspace Wiki
→How It Works
| Line 4: | Line 4: | ||
The shitcam has been placed in a strategic, discrete location in the hackspace to capture people when sitting down to use the facilities. Instead of sending out a video stream, the raw data is continuously processed by a Processing sketch which uses [http://ubaa.net/shared/processing/opencv/ OpenCV] to implement face detection. When a face is detected a snapshot is uploaded to twitpic, using the [http://processing.org/discourse/yabb2/YaBB.pl?num=1258068743 twitpic4p] library. | The shitcam has been placed in a strategic, discrete location in the hackspace to capture people when sitting down to use the facilities. Instead of sending out a video stream, the raw data is continuously processed by a Processing sketch which uses [http://ubaa.net/shared/processing/opencv/ OpenCV] to implement face detection. When a face is detected a snapshot is uploaded to twitpic, using the [http://processing.org/discourse/yabb2/YaBB.pl?num=1258068743 twitpic4p] library. | ||
Before being twittered, the image is cropped to the approximate area of the detected face. This is important in order to respect people's privacy, and to avoid accidental prickpics, twatpics et c. Under testing I've found that extraneous bodyparts have only appeared when they actually look like | Before being twittered, the image is cropped to the approximate area of the detected face. This is important in order to respect people's privacy, and to avoid accidental prickpics, twatpics et c. Under testing I've found that extraneous bodyparts have only appeared when they actually look like faces. | ||
== Access == | == Access == | ||
