学生のGALAJIT, KasornさんがiSAI-NLP 2018にてBest Paper Awardを受賞
iSAI-NLP（the International joint Symposium on Artificial Intelligence and Natural Language Processing 旧名称：SNLP(the Symposium on Natural Language Processing)）は、タイを発祥として1993年から開催されている自然言語処理と人工知能に特化した国際会議のシリーズの1つです。今回のiSAI-NLP 2018は、13回目の開催となり、2018年11月15日～17日にかけてタイのパタヤで開催されました。
Speech Watermarking Technique Based on Singular Spectrum Analysis and Automatic Parameter Estimation using Differential Evolution for Tampering Detection
Kasorn Galajit, Mongkonchai Intarauksorn, Jessada Karnjana, Pakinee Aimmanee, Masashi Unoki
With the aim of detecting a tampering in a speech signal, this paper proposed a singular spectrum analysis-based watermarking scheme with automatic parameter estimation using differential evolution. At the sender side, the singular spectrum of an original speech signal is modified to embed the watermark bit, and differential evolution is used to select part of singular spectrum to be modified in order to balance between the robustness of the scheme and the sound quality of the watermarked signal. At the receive side, the watermark is extracted and compared with the original watermark for tampering detection. The experimental results show that the analysis of an extracted watermark can be used to identify the tampering in both location and the degree of the attacks.
It is a great honor to receive the best paper award of ISAI 2018. I am thankful to organizing committee to consider my research for the best paper prize. Actually, the person who really deserves credit on this is not only me but also all of my co-authors. I would like to convey my humble gratitude to my advisor, Professor Masashi Unoki, for his continuous support. Also I am grateful to Dr. Jessada Karnjana and Dr. Pakinee Aimmanee for their great guidance and encouragement. Last but not least, Mongkonchai Intarauksorn who also received the best presentation award for his presentation on this papers. The best paper award is giving me a motivation to achieve a quality research in the future.