The ASR test set will be delivered by March 8, 2018 at 8:30 am via email.
Evaluation data: The test set is composed of continuous .wav files of news speech for a total duration of two hours, without any information on the sentence segmentation. The speech was recorded in a non-noisy environment. The speech data is available in three dialects: Northern, Southern and Central with respectively proportion of 50%, 40% and 10%.
Evaluation measure: WER (Word Error Rate)
Result format: The result should be saved in separated plain text files corresponding to input files, spoken word, lower case, no punctuation, UTF-8. For example:
filename-1 ngày một tháng sáu ủy ban nhân dân tỉnh bà rịa vũng tàu
filename-2 kính chào quý vị khán giả
...
filename-n hẹn gặp lại các bạn
Result submission: The log file of the system should be submitted to M. Nguyễn Văn Huy (huynguyen at tnut dot edu in vn) before 6:00 pm on March 10, 2018. It is possible for each team for submitting several times the result before the deadline, but only the last submission will be taken into account.
A technical report of each participant's system should be submitted before 6:00 pm on March 15, 2018.