Mercurial > hg > nsaunier > traffic-intelligence

comparison classifier.cfg @ 1243:88eedf79f16a

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
correct classifier.cfg and info classification
author Nicolas Saunier <nicolas.saunier@polymtl.ca>
date Wed, 07 Feb 2024 11:53:00 -0500
parents ab4c72b9475c
children
comparison
equal deleted inserted replaced
1242:4cd8ace3552f 1243:88eedf79f16a
1 # filename of the general ped/cyc/veh SVM classifier 1 # filename of the general ped/cyc/veh SVM classifier
2 pbv-svm-filename = modelPBV.xml 2 pbv-svm-filename = modelPBV.xml
3 # filename of the cyc/veh SVM classifier 3 # filename of the cyc/veh SVM classifier
4 bv-svm-filename = modelBV.xml 4 bv-svm-filename = modelBV.xml
5 # filename of a Ultralytics-compatible model, eg Yolov8 5 # filename of a Ultralytics-compatible model, eg Yolov8
6 dl-filename = 6 # if the filename is valid, it will be used, otherwise the SVMs
7 dl-filename = yolov8x.pt
7 # percent increase of the max of width and height of the bounding box of features extracted for classification 8 # percent increase of the max of width and height of the bounding box of features extracted for classification
8 percent-increase-crop = 0.2 9 percent-increase-crop = 0.2
9 # min number of pixels in cropped image to classify by SVM 10 # min number of pixels in cropped image to classify by SVM
10 min-npixels-crop = 800 11 min-npixels-crop = 800
11 # square size to resize image crops for HoG computation 12 # square size to resize image crops for HoG computation
16 hog-npixels-cell = 8 17 hog-npixels-cell = 8
17 # number of cells per block for HoG computation 18 # number of cells per block for HoG computation
18 hog-ncells-block = 2 19 hog-ncells-block = 2
19 # block normalization method (L1, L1-sqrt, L2, L2-Hys) 20 # block normalization method (L1, L1-sqrt, L2, L2-Hys)
20 hog-block-norm = L1-sqrt 21 hog-block-norm = L1-sqrt
22 # object confidence threshold for detection
23 confidence = 0.25
21 # method to aggregate road user speed: mean, median or any (per)centile 24 # method to aggregate road user speed: mean, median or any (per)centile
22 speed-aggregation-method = median 25 speed-aggregation-method = median
23 # number of frames to ignore at both ends of a series (noisy) 26 # number of frames to ignore at both ends of a series (noisy)
24 nframes-ignore-at-ends = 2 27 nframes-ignore-at-ends = 2
25 # centile for the speed aggregation, if centile is chosen 28 # centile for the speed aggregation, if centile is chosen