Skip to content

Latest commit

 

History

History
144 lines (107 loc) · 4.95 KB

File metadata and controls

144 lines (107 loc) · 4.95 KB

Dataset Preparation

Kinetics

The Kinetics Dataset could be downloaded via the code released by ActivityNet:

  1. Download the videos via the official scripts.

  2. After all the videos were downloaded, resize the video to the short edge size of 256, then prepare the csv files for training, validation, and testing set as train.csv, val.csv, test.csv. The format of the csv file is:

path_to_video_1 label_1
path_to_video_2 label_2
path_to_video_3 label_3
...
path_to_video_N label_N

All the Kinetics models in the Model Zoo are trained and tested with the same data as Non-local Network. For dataset specific issues, please reach out to the dataset provider.

AVA

The AVA Dataset could be downloaded from the official site

We followed the same downloading and preprocessing procedure as the Long-Term Feature Banks for Detailed Video Understanding do.

You could follow these steps to download and preprocess the data:

  1. Download videos
DATA_DIR="../../data/ava/videos"

if [[ ! -d "${DATA_DIR}" ]]; then
  echo "${DATA_DIR} doesn't exist. Creating it.";
  mkdir -p ${DATA_DIR}
fi

wget https://s3.amazonaws.com/ava-dataset/annotations/ava_file_names_trainval_v2.1.txt

for line in $(cat ava_file_names_trainval_v2.1.txt)
do
  wget https://s3.amazonaws.com/ava-dataset/trainval/$line -P ${DATA_DIR}
done
  1. Cut each video from its 15th to 30th minute
IN_DATA_DIR="../../data/ava/videos"
OUT_DATA_DIR="../../data/ava/videos_15min"

if [[ ! -d "${OUT_DATA_DIR}" ]]; then
  echo "${OUT_DATA_DIR} doesn't exist. Creating it.";
  mkdir -p ${OUT_DATA_DIR}
fi

for video in $(ls -A1 -U ${IN_DATA_DIR}/*)
do
  out_name="${OUT_DATA_DIR}/${video##*/}"
  if [ ! -f "${out_name}" ]; then
    ffmpeg -ss 900 -t 901 -i "${video}" "${out_name}"
  fi
done
  1. Extract frames
IN_DATA_DIR="../../data/ava/videos_15min"
OUT_DATA_DIR="../../data/ava/frames"

if [[ ! -d "${OUT_DATA_DIR}" ]]; then
  echo "${OUT_DATA_DIR} doesn't exist. Creating it.";
  mkdir -p ${OUT_DATA_DIR}
fi

for video in $(ls -A1 -U ${IN_DATA_DIR}/*)
do
  video_name=${video##*/}

  if [[ $video_name = *".webm" ]]; then
    video_name=${video_name::-5}
  else
    video_name=${video_name::-4}
  fi

  out_video_dir=${OUT_DATA_DIR}/${video_name}/
  mkdir -p "${out_video_dir}"

  out_name="${out_video_dir}/${video_name}_%06d.jpg"

  ffmpeg -i "${video}" -r 30 -q:v 1 "${out_name}"
done
  1. Download annotations
DATA_DIR="../../data/ava/annotations"

if [[ ! -d "${DATA_DIR}" ]]; then
  echo "${DATA_DIR} doesn't exist. Creating it.";
  mkdir -p ${DATA_DIR}
fi

wget https://research.google.com/ava/download/ava_train_v2.1.csv -P ${DATA_DIR}
wget https://research.google.com/ava/download/ava_val_v2.1.csv -P ${DATA_DIR}
wget https://research.google.com/ava/download/ava_action_list_v2.1_for_activitynet_2018.pbtxt -P ${DATA_DIR}
wget https://research.google.com/ava/download/ava_train_excluded_timestamps_v2.1.csv -P ${DATA_DIR}
wget https://research.google.com/ava/download/ava_val_excluded_timestamps_v2.1.csv -P ${DATA_DIR}
  1. Download "frame lists" (train, val) and put them in the frame_lists folder (see structure above).

  2. Download person boxes (train, val, test) and put them in the annotations folder (see structure above). If you prefer to use your own person detector, please see details in here.

Download the ava dataset with the following structure:

ava
|_ frames
|  |_ [video name 0]
|  |  |_ [video name 0]_000001.jpg
|  |  |_ [video name 0]_000002.jpg
|  |  |_ ...
|  |_ [video name 1]
|     |_ [video name 1]_000001.jpg
|     |_ [video name 1]_000002.jpg
|     |_ ...
|_ frame_lists
|  |_ train.csv
|  |_ val.csv
|_ annotations
   |_ [official AVA annotation files]
   |_ ava_train_predicted_boxes.csv
   |_ ava_val_predicted_boxes.csv

You could also replace the v2.1 by v2.2 if you need the AVA v2.2 annotation. You can also download some pre-prepared annotations from here.