In testing the classifiers produced, I ran it initially against the training set that was used to create the classifier, as seen in the previous post. So now we are going to start running tests on new image sets, to see how it truly performs on images it has never “seen” before.
New Image Nov. 24, 2011
The below hit rate was achieved when running the new classifier against the above single image rotated by one degree for 360 degrees.
Hit rate ≈ 90%
+================================+======+======+======+ | File Name | Hits |Missed| False| +================================+======+======+======+ | 20111124_002656_1024_0193.jpg| 1| 0| 21| +--------------------------------+------+------+------+ | 1-DEG-24_002656_1024_0193.jpg| 1| 0| 28| +--------------------------------+------+------+------+ [...] +--------------------------------+------+------+------+ | 358-DEG-24_002656_1024_0193.jpg| 1| 0| 20| +--------------------------------+------+------+------+ | 359-DEG-24_002656_1024_0193.jpg| 1| 0| 25| +--------------------------------+------+------+------+ | Total| 324| 36| 10341| +================================+======+======+======+
Tags: haar training, performance, testing outcomes, trials
OK so we get a fairly good hit rate but we have the high false positives which is probably due to the high sensitivity problem we discussed…
Lets keep going as planned from yesterday! Keep up the good work!