why we should create .vec files to train harr cascade classifier.
Hi, I am not understanding why we are creating .vec files during positive samples creation using opencv_createsamples. Why it is required and how the image data is stored in it (any format?). Please reply i am a beginner in opencv, thanks in advance.
Read: https://docs.opencv.org/master/dc/d88...
I am able to create samples and generate .vec files and also able to merge all .vec files into one file using mergevec.py , but i want to know why we are creating vector files for each image sample?.
@Akshata7, please read the tutorial carefully. The official one is NEVER using a
mergevec.py
program, because that described approach in several tutorials simply does not yield decent results. IF you want to know why, have a read of chapter 5 in the book OpenCV 3 Blueprints, which discusses the whole pipeline in detail. It has also been explained over thousand times on this forum, that using the artificially modified data does not help to build robuster detectors.