mumuyeye / fade Goto Github PK

Our dataset, annotation instructions, dataset documentation and intended uses are published and you can download our dataset from here

1. download the dataset from here
2. unzip the dataset and put it in the dataset folder
3. run python train.py to train the model with the default parameters defined in train.py. the usage is as same as original yolov5 method's usage. you can also change the parameters by yourself.

1. create a file named test.txt, and put the path of the test videos in it
2. run python test.py --val test.txt
3. the results will be saved in the video folder

CPU: Intel(R) Xeon(R) Gold 5117 CPU @ 2.00GHz GPU: GeForce RTX 3090

source code: BSUV-NET-2.0

1. python: 3.8.8
2. pytorch: 1.8.1
3. opencv: 4.0.1

1. network: unetvgg16
2. inp_size: 600
3. empty_bg: manual
4. recent_bg: None
5. seg_ch: None
6. lr: 0.0001
7. weight_decay: 0.01
8. opt: adam
9. aug_ioa: None

source code: MOBILE-VOD

1. python: 3.8.12
2. pytorch: 1.11.0
3. opencv-python: 4.5.5.64

1. network: mvod_lstm1
2. lr: 0.003
3. momentum: 0.9
4. weight_decay: 0.0005
5. gamma: 0.1

Run python precision_recall_tro.py --algo <k> for <k> = MOG2, KNN, ... to compute the results for each fold. This code will save the results to result.txt.

1. python: 3.8.8
2. opencv-contrib-python: 4.5.5.64

opencv::createBackgroundSubtractorMOG2()

1. history: 300
2. varThreshold: 100
3. detectShadow: false

opencv::createBackgroundSubtractorKNN()

1. history: 300
2. varThreshold: 100
3. detectShadow: false

opencv::bgsegm::createBackgroundSubtractorMOG()

• default

opencv::bgsegm::createBackgroundSubtractorCNT()

1. setIsParallel
2. setUseHistory
3. setMinPixelStability(1)
4. setMaxPixelStability(4)

cv::bgsegm::BackgroundSubtractorGMG

1. setNumFrames(5)
2. setUpdateBackgroundModel(True)

cv::bgsegm::BackgroundSubtractorGSOC

1. history: 300
2. varThreshold: 100

cv::bgsegm::BackgroundSubtractorLSBP

• default

Run python vibe.py to compute the results for each fold. This code will save the results to result.txt.

1. num_sam: 20
2. min_match: 2
3. radiu: 20
4. rand_sa: 16