This shows you the differences between two versions of the page.
Both sides previous revision Previous revision | |||
space:kitchen:foodcam [2013/12/05 13:55] jkuroda |
space:kitchen:foodcam [2013/12/05 14:14] (current) apanda |
||
---|---|---|---|
Line 3: | Line 3: | ||
The [[https://radlab.cs.berkeley.edu/foodcam|RAD Lab Kitchen Free Food Camera]] allows one to view what is currently on the Kitchen Table without having to always visit the kitchen (and then become engaged in loud conversations in the kitchen). | The [[https://radlab.cs.berkeley.edu/foodcam|RAD Lab Kitchen Free Food Camera]] allows one to view what is currently on the Kitchen Table without having to always visit the kitchen (and then become engaged in loud conversations in the kitchen). | ||
- | Panda and Kay are currently running a computer vision experiment to detect the presence of food on the table and, eventually, identify the type of food via Deep Learning™ | + | Panda and Kay are currently running a computer vision experiment to detect the presence of food on the table and, eventually, identify the type of food via the use of a simple SVM thing. We hope one day someone smarter will use Deep Learning to improve on these results (note this meets all three of A, M and P). |
The camera is positioned such that only the table itself is within view so that people are not accidentally viewed by the camera ((We currently use an ad-hoc Cone of Shame™ for this but will later acquire a better camera with a zoom feature. If you choose to place yourself into this limited field of view, you will likely be caught on camera)) | The camera is positioned such that only the table itself is within view so that people are not accidentally viewed by the camera ((We currently use an ad-hoc Cone of Shame™ for this but will later acquire a better camera with a zoom feature. If you choose to place yourself into this limited field of view, you will likely be caught on camera)) | ||
It is limited to members of the EECS community via passwd/login. | It is limited to members of the EECS community via passwd/login. |