Wangyida's profile - activity

This could solved by checking the rootpath and make it empty if such exception occurs.

I thought some dense feature is useful for 2D detection and object classification. Such method could be DENSE SIFT, hidimLBP and advanced learnd features, I thought these features can be adapted for many 2D application if such work is worthful

The octave is usually been set as 8 and there are also different scales of the original picture, this could be regard as filtered be different Gaussian filters. The patch I saw most is set as 44 or 22, each contains 44--66 pixals.

Maybe you could train the caegory which isn't digt the 11th catogory, or use a two step SVM the filter them

I would like to use it in local linea svm, but it's incovenient without precomputed kernel. How to solve the problem?

When run CmFile::GetName(CStr& path), The path just doesn't exist in my hard drive and the debugging process may be just failed without accurate description. First-chance exception at 0x00007FF827D4606C in Objectness.exe: Microsoft C++ exception: cv::Exception at memory location 0x00000054D845E930. Unhandled exception at 0x00007FF827D4606C in Objectness.exe: Microsoft C++ exception: cv::Exception at memory location 0x00000054D845E930.