@article{oai:nagasaki-u.repo.nii.ac.jp:00014412,
 author = {黒田, 英夫 and 藤村, 誠 and 下田, 裕 and マワンサ, デニス シレッシュ},
 issue = {40},
 journal = {長崎大学工学部研究報告, Reports of the Faculty of Engineering, Nagasaki University},
 month = {Jan},
 note = {In the international standard video coding specification H. 261 for videophone systems. DCT (Discrete Cosine Transform) is used for the interframe difference signals. If the used DCT algorithms are different each other, FDCT-IDCT mismatch may occur and cause the introduction of mismatch noise in the frame memorys of sender and reciever. The leaky prediction is useful for removing the mismatch noise although it causes decreasing of insignificant blocks or deteriorating of picture quality on static area. In this paper, the leaky interframe prediction coding system is proposed. The system however prevents insignificant blocks decreasing by raising up the threshold value to decide between insignificant block and significant block. Moreover, by dividing the decoded signals by leaky prediction coefficient, the decoded picture quality is improved for displaying., 長崎大学工学部研究報告, 23(40), pp.41-48; 1993},
 pages = {41--48},
 title = {リーク予測による異種DCTフレーム間符号化方式の整合化},
 volume = {23},
 year = {1993}
}