What do you think is taking more juice from the cpu? finding the signs in those high resolution images? or doing the image recognition?
I see that you are using a pretty powerful cpu for this (3Ghz quad).
I find it amazing that small digital cameras nowadays run super fast detection of multiple faces, but I doubt they use such a powerful cpu as you are using. I wonder how they manage to do that. Using a specialized processor, like a DSP I guess?
Can you give me any hint about how you defined the Noise class? For example in one case scenario I used about 2500 normalized images of the 30 speed limit sign and about 1800 images labeled as Noise. In training/testing the SVM classifier I got over 90% accuracy but when testing on real video, almost everything gets labeled as Noise.
Are you doing traffic sign detection or recognition?because your system does not specify the type of road sign,it just detect and show it.
natfeh1982 7 months ago
What do you think is taking more juice from the cpu? finding the signs in those high resolution images? or doing the image recognition?
I see that you are using a pretty powerful cpu for this (3Ghz quad).
I find it amazing that small digital cameras nowadays run super fast detection of multiple faces, but I doubt they use such a powerful cpu as you are using. I wonder how they manage to do that. Using a specialized processor, like a DSP I guess?
echelecopao 8 months ago
Which technical language have you used to implement this project: C, Visual C++ or Matlab?
bmt1903 8 months ago
Hi Which camera did you choose?
warisshaw1 1 year ago
Hi Amazing work!!
Which camera you used ?
warisshaw1 1 year ago
Comment removed
adyandon 1 year ago
Comment removed
adyandon 1 year ago
Hi moonhwiLee!
Can you give me any hint about how you defined the Noise class? For example in one case scenario I used about 2500 normalized images of the 30 speed limit sign and about 1800 images labeled as Noise. In training/testing the SVM classifier I got over 90% accuracy but when testing on real video, almost everything gets labeled as Noise.
Thanks in advance!
adyandon 1 year ago
great.....what technique are u using?
template matching?
ankit123ful 1 year ago
@ankit123ful
I use color-based methods for ROI extraction.
NN+SVM is used to recognize road sign.
moonhwiLee 1 year ago
@moonhwiLee
What does NN+SVM mean? Can you please give a little more details? Thanks!
adyandon 1 year ago
@adyandon
Please find these algorithm:-)
NN : Neural Network
SVM : Support Vector Machine
moonhwiLee 1 year ago
good work, well done
omalleyronan 1 year ago
Great Work!
s2zat 1 year ago
Impressive demo :-)
sherv81 1 year ago