2014-03-23 10:23:16 -0600 | received badge | ● Supporter (source) |
2013-10-25 20:17:20 -0600 | answered a question | Multiple simultaneous logo/target recognition/reporting Option A: train cascades for each logo. Then use object detection. Option B: use 2d features keypoints and descriptors, then find using matching (Flan or brute force) Extraction and emailing is on you after that. |
2013-10-25 20:10:46 -0600 | received badge | ● Editor (source) |
2013-10-25 20:09:45 -0600 | answered a question | opencv_traincascade training time/stuck When I tried to train with a buffer that large train took 8GB+ ram. When its needs exceeded my available ram the whole process slowed way down. I assume it started using my hdd as memory overflow. Try 512mb. Also monitor your CPU and ram usage. What is it at when it seems to hang. |
2013-10-25 20:03:27 -0600 | answered a question | How does number of training stage affect the performance of cascade file? With 30 stages you might have over trained your classifier. I'm running into a similar problem. Did you crop positives from the video you are testing with? Finding the perfect parameters for cascade training is the key. For my project I have a trial journal where I note all parameters and the result. Trial and error takes a long time when training takes a few hours. |