US 12,333,440 B2
Signal processing apparatus, signal processing system, signal processing method, and computer program product
Tenta Sasaya, Ota (JP); and Takashi Watanabe, Yokohama (JP)
Assigned to KABUSHIKI KAISHA TOSHIBA, Minato-ku (JP)
Filed by KABUSHIKI KAISHA TOSHIBA, Minato-ku (JP)
Filed on Feb. 24, 2020, as Appl. No. 16/798,534.
Claims priority of application No. 2019-103658 (JP), filed on Jun. 3, 2019.
Prior Publication US 2020/0380281 A1, Dec. 3, 2020
Int. Cl. G06N 3/084 (2023.01); G06N 3/04 (2023.01); G06N 3/08 (2023.01); G06V 10/30 (2022.01); G06V 10/764 (2022.01); G06V 10/82 (2022.01); G06V 10/98 (2022.01); G10L 25/21 (2013.01); G10L 25/51 (2013.01)
CPC G06N 3/084 (2013.01) [G06N 3/08 (2013.01); G06V 10/30 (2022.01); G06V 10/764 (2022.01); G06V 10/82 (2022.01); G06V 10/98 (2022.01); G06N 3/04 (2013.01); G10L 25/21 (2013.01); G10L 25/51 (2013.01)] 12 Claims
OG exemplary drawing
 
1. A signal processing apparatus, comprising:
processing circuitry configured to function as:
a signal processing unit configured to perform first signal processing on an input signal in which a second signal is superimposed on a first signal, the first signal processing being performed using a plurality of parameters, and to output a first restoration signal obtained by restoring the first signal using the first signal processing;
a first signal adjustment unit configured to receive the first signal and the first restoration signal as inputs, multiply both of the first signal and the first restoration signal that are input by respective weighting coefficients, and output a corrected first signal and a corrected first restoration signal; and
an update unit configured to update the plurality of parameters using an error between the corrected first signal and the corrected first restoration signal output from the first adjustment unit, wherein the weighting coefficients are expressed by a matrix with coefficients arranged on a main diagonal, and the respective weighting coefficients used by the first signal adjustment unit are set to make the error larger than a signal strength of the second signal.