I have created nine classifiers (numbers 1-9) using the opencv_traincascade utility. Now I want to use them to detect some numbers, the problem is: when I pass an image with a number, more than one classifier detect his number (e.g. numbers 1 and 7 because they look similar), so is there a way to find out the best matching? Thanks in advance
Asked
Active
Viewed 429 times
0
-
Please share your code... – Shiva Mar 03 '16 at 10:45
-
You don't need code. I just have nine CascadeClassifier trained with number images. I only want to know if there is a method to find out the best matching among these classifiers. – Max Mar 03 '16 at 10:56
-
OpenCv is written in C++ and you have tagged this question in Java, how are you calling opencv functions, are you using a library like JavaCV – Shiva Mar 03 '16 at 11:28
-
OpenCV supports Java since 2013... – Max Mar 03 '16 at 11:42
-
Sorry I was a bit outdated on that ! – Shiva Mar 03 '16 at 11:54
-
you are abusing an object*detection* mechanism for a *classification* problem. this won't work. – berak Mar 03 '16 at 16:14