Skip to content

Latest commit

 

History

History
122 lines (107 loc) · 5.95 KB

TrafficCAM-download.md

File metadata and controls

122 lines (107 loc) · 5.95 KB
Raw

The dataset (TrafficCAM) can be downloaded from the links below:

The fully annotated data can be downloaded at:
Fully_annotate.zip https://drive.google.com/file/d/1h4oUqECDF05vSYMkYgz0aUc_RRzHlh3e/view?usp=sharing

The first frame annotated data is divided into 3 partitions:
Partition 1:
FirstFrame_annotate-1.zip https://drive.google.com/file/d/1lYz_aBX-4Ygj-tbUkHvsaPB3vxOIvGBA/view?usp=sharing
Partition 2:
FirstFrame_annotate-2.zip https://drive.google.com/file/d/1TmN07tdC5NYpcIPI5tmhOrU1XfK7Yu6V/view?usp=share_link
Partition 3:
FirstFrame_annotate-3.zip https://drive.google.com/file/d/1cKOpDQ-NvqTxIETXYcDOv_Ojc3UZm2Qq/view?usp=sharing

The data structure for using the dataset as suggested in benchmark can be viewed below:

  ├── TrafficCAM
    ├── raw
      ├── FirstFrame_annotate
        ├── [Video1_ID]
        :  ├── frame0.json   	# annotation of frame0 (under FirstFrame_annotate only the 1st frame is annotated)
        :  ├── frame0.jpg	# raw frame0
        :  ├── frame2.jpg	# raw frame2
        :  :
        :  :	### 30 raw frames per video 
        :  :	### NOTE: in video "BLR_1651662434.7875617", only 23 raw frames: missed frame6,8,50,52,54,56,58.
        :	 :		        
        :  └── frame58.jpg	# raw frame58
        :
        :	### 2024 videos with only first frame annotated in total
        :
        └──  [Video2024_ID] 	 
      └── Fully_annotate
        ├── [Video1_ID]
        :  ├── frame0.json   	# annotation of frame0
        :  :
        :  :	 ### 30 annotated frames in total (under Fully_annotate all frames annotated)
        :  :
        :  ├── frame58.json	# annotation of frame58
        :  ├── frame2.jpg	# raw frame2
        :  :
        :  :	 ### 30 raw frames in total
        :  :
        :  └── frame58.jpg	# raw frame58
        :	
        :	### 78 videos with all frames annotated in total
        :	
        └──  [Video78_ID]
   
    └── splits          # official splits
      ├── benchmark     # training set, (validation set), and test set used in benchmark
        ├── train
            ├── semi-supervised
                ├── 1_27
                    ├── train_labelled.txt
                    └── train_unlabelled.txt
                ├── 1_54
                  ├── train_labelled.txt
                  └── train_unlabelled.txt
                ├── 1_108
                  ├── train_labelled.txt
                  └── train_unlabelled.txt
                └──1_216
                  ├── train_labelled.txt
                  └── train_unlabelled.txt
            └── supervised
                └──  train.txt
        ├── val.txt
        └── test.txt
      └── leave-out.txt			# 12 fully annotated videos are left out for other use

The benchmarked split in semi-supervised training can be found in:

splits.zip https://drive.google.com/file/d/1GcQ3J56kU-6TwRqRjzvI28-OF6QUdHeR/view?usp=sharing

Instruction for downloading and assembling the data as benchmarked is as below through commend lines:

Create a clean folder

mkdir TrafficCAM
cd TrafficCAM
mkdir raw
cd raw

Download the data

wget --load-cookies /tmp/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id=1lYz_aBX-4Ygj-tbUkHvsaPB3vxOIvGBA' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\1\n/p')&id=1lYz_aBX-4Ygj-tbUkHvsaPB3vxOIvGBA" -O FirstFrame_annotate-1.zip && rm -rf /tmp/cookies.txt 
wget --load-cookies /tmp/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id=1TmN07tdC5NYpcIPI5tmhOrU1XfK7Yu6V' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\1\n/p')&id=1TmN07tdC5NYpcIPI5tmhOrU1XfK7Yu6V" -O FirstFrame_annotate-2.zip && rm -rf /tmp/cookies.txt
wget --load-cookies /tmp/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id=1cKOpDQ-NvqTxIETXYcDOv_Ojc3UZm2Qq' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\1\n/p')&id=1cKOpDQ-NvqTxIETXYcDOv_Ojc3UZm2Qq" -O FirstFrame_annotate-3.zip && rm -rf /tmp/cookies.txt
wget --load-cookies /tmp/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id=1h4oUqECDF05vSYMkYgz0aUc_RRzHlh3e' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\1\n/p')&id=1h4oUqECDF05vSYMkYgz0aUc_RRzHlh3e" -O Fully_annotate.zip && rm -rf /tmp/cookies.txt

Unzip the folders

sudo apt install unzip
unzip FirstFrame_annotate-1.zip
unzip FirstFrame_annotate-2.zip
unzip FirstFrame_annotate-3.zip
unzip Fully_annotate.zip

Assemble the first frame annotated data:

mkdir FirstFrame_annotate
mv FirstFrame_annotate-1/* FirstFrame_annotate-2/* FirstFrame_annotate-3/* FirstFrame_annotate/

Download the benchmarked splits

cd ../
wget --load-cookies /tmp/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id=1GcQ3J56kU-6TwRqRjzvI28-OF6QUdHeR' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\1\n/p')&id=1GcQ3J56kU-6TwRqRjzvI28-OF6QUdHeR" -O splits.zip && rm -rf /tmp/cookies.txt 
unzip splits.zip

Delete all preparation files

rm raw/*.zip
rm -r raw/*-*
rm *.zip
rm -r __MACOSX