Codes for retrieving data in the corpus SAIT-EMA
from SAIT_EMA_Reader import SAIT_RawEMA_Reader
EMA_Reader = SAIT_RawEMA_Reader(raw_EMA_dir, speaker_name)raw_EMA_dir is the absolute path of pos diretory (caution, not the rawpos directory) of the current speaker speaker_name.
All possible speaker names: F001, F002, F003, F004, F005, M001, M002, M003, M004, M005, A_M001, A_F001, L_F001, L_F002, L_F003, L_M001, L_M002, L_M003.
The term std_EMA refers to a numpy.ndarray with the shape of (num_frames, 12). 12 stands for the x, y coordinates of the six sensors: tt, tb, td, lj, ul, ll (Tongue Tip, Tongue Body, Tongue Dorsum, Lower Jaw, Upper Lip, Lower Lip).
Each speaker has exactly 962 std_EMAs.
The term nose_EMA refers to a numpy.ndarray with the shape of (num_frames, 2). It is the x, y coordinates of the NOSE sensor.
The term "index" refers to the filename without extension. For example, for the file 0001.pos, its index is 0001. An "index list" is a Python list whose elements are all indices. For example, ["0001", "0002", "0900"] is an index list.
Extracting std_EMA by providing the absolute path of the .pos file.
std_EMA = EMA_Reader.get_std_EMA(path)Extracting std_EMA by providing its index.
std_EMA = EMA_Reader.get_std_EMA_by_index(index)Get a list of std_EMA by providing the corresponding index list.
If index list is None, return a list containing all std_EMA of the current speaker.
std_EMA_list = EMA_Reader.get_std_EMA_list_by_index_list(index_list)Get a random std_EMA.
std_EMA = EMA_Reader.get_randn_std_EMA()Extracting nose_EMA by providing the absolute path of the .pos file.
nose_EMA = EMA_Reader.get_nose_EMA_by_path(path)Extracting nose_EMA by providing its index.
nose_EMA = EMA_Reader.get_nose_EMA_by_index(index)Get a list of nose_EMA by providing the corresponding index list.
If index list is None, return a list containing all nose_EMA of the current speaker.
nose_EMA_list = EMA_Reader.get_nose_EMA_list_by_index_list(index_list)