Skip to content

Latest commit

 

History

History
45 lines (28 loc) · 1.62 KB

readme.md

File metadata and controls

45 lines (28 loc) · 1.62 KB

Isolated Sign Language Recognition using Conv3D

This folder contains the training, finetuning code for the following modalities.

Pretrained models

Pretrained models can be downloaded via Google Drive. Those pretrained models can be used to reproduce our final results using the testing code.

Requirement

See requirements.txt

RGB Frames

The RGB frames modality can be trained, finetuned and tested using the following commands in Conv3D/ folder.

python Sign_Isolated_Conv3D_clip.py

python Sign_Isolated_Conv3D_clip_finetune.py

python Sign_Isolated_Conv3D_clip_test.py

Flow Color

The Flow Color modality can be trained, finetuned and tested using the following commands

python Sign_Isolated_Conv3D_flow_clip.py

python Sign_Isolated_Conv3D_flow_clip_funtine.py

python Sign_Isolated_Conv3D_flow_clip_test.py

HHA

The HHA modality can be trained, finetuned and tested using the following commands

python Sign_Isolated_Conv3D_hha_clip_mask.py

python Sign_Isolated_Conv3D_hha_clip_mask_finetune.py

python Sign_Isolated_Conv3D_hha_clip_mask_test.py

Flow Depth

The Flow Depth modality can be trained, finetuned and tested using the following commands

python Sign_Isolated_Conv3D_depth_flow_clip.py

python Sign_Isolated_Conv3D_depth_flow_clip_finetune.py

python Sign_Isolated_Conv3D_depth_flow_clip_test.py

Ensemble

The results .pkl files of the above modalities will be saved in results/ folder. Please rename and copy them to ../ensemble/ folder for model ensemble.