traffic sign detection with HOG feature and SVM model
see details in my blog: https://blog.csdn.net/sinat_34474705/article/details/80502789
test_image.py: detect the signs on an image.
test_video.py: detect the signs on a video.
test_dir.py: detec the images in a directory (save tthe result in a txt file).
evaluation.py: compute the precision and recall based on the detect result txt file, which needs the label txt file (generated by tools/preprocess_traffic_signs.py)