This repository contains the code for the following paper: Kim, Jungi and Patricia O'Neill-Brown. “Improving American Sign Language Recognition with Synthetic Data.” MTSummit (2019). https://www.aclweb.org/anthology/W19-6615/
If you use this for your research work, please cite the following paper:
@inproceedings{
title = "Improving {A}merican Sign Language Recognition with Synthetic Data",
author = "Kim, Jungi and
O{'}Neill-Brown, Patricia",
booktitle = "Proceedings of Machine Translation Summit XVII Volume 1: Research Track",
month = aug,
year = "2019",
address = "Dublin, Ireland",
publisher = "European Association for Machine Translation",
url = "https://www.aclweb.org/anthology/W19-6615",
pages = "151--161",
}
We used an docker image based on the following dockerfile: https://gist.github.com/moiseevigor/11c02c694fc0c22fccd59521793aeaa6
Each ASL video was processed by the following docker command:
nvidia-docker run \
-v ${LOCAL_DIR}:/workspace \
--rm -it openpose1.4:Dockerfile \
bash -c "\
CUDA_VISIBLE_DEVICES=0 \
/opt/openpose-master/build/examples/openpose/openpose.bin \
-model_folder /opt/openpose-master/models/ \
-model_pose BODY_25 \
-face true \
-face_render 1 \
-hand true \
-hand_render 1 \
-hand_tracking true \
-display 0 \
-number_people_max 1 \
-video /workspace/${INPUT_VIDEO} \
-write_images /workspace/${OUTPUT_DIR} \
-write_json /workspace/${OUTPUT_DIR} \
-write_video /workspace/${OUTPUT_DIR}/output.avi"
where LOCAL_DIR
is the host directory containing a video file INPUT_VIDEO
,
and OUTPUT_DIR
is where the OpenPose analysis output is created.
DeepHand model and the code to evaluate the model is taken from https://github.com/neccam/TF-DeepHand.
./generateFrames.sh ${LOCAL_DIR}/${INPUT_VIDEO} ${OUTPUT_DIR}
python GenerateASLRFeatures.py -i ${OUTPUT_DIR}
Features are stored as a matlab file with name ${OUTPUT_DIR}.mat