Intra-instrument channel workable, optical-resolution photoacoustic and ultrasonic mini-probe system for gastrointestinal endoscopy

Deep-Learning-Based Algorithm for the Removal of Electromagnetic Interference Noise in Photoacoustic Endoscopic Image Processing

“…In the case of photoacoustic endoscopy (PAE) or intravascular photoacoustic (IVPA) imaging applications [ 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 ], related issues appear more frequently and seriously because it is a common hardware situation that the US transducer for signal detection and the associated preamplifier for the amplification of the detected signal are placed separately, with one at the distal end of the flexible probe section and the other with the proximal driving unit, at a distance of more than ~1.5 m, as the space allowed for installing a preamplifier circuit at the distal scanning tip is very limited. Moreover, due to the same limited space problem, it is not easy to apply proper shielding material to the connection path (i.e., the flexible probe section), which is located between the two parts mentioned [ 7 ].…”

“…Moreover, due to the same limited space problem, it is not easy to apply proper shielding material to the connection path (i.e., the flexible probe section), which is located between the two parts mentioned [ 7 ]. Consequently, it has been recognized that it is very difficult to implement a perfect hardware setup in PAE or IVPA that is not affected by interference noise at all, especially where the system is embodied at a laboratory level, and the problem related to the involvement of noise has been mentioned frequently in many reports [ 19 , 20 , 21 , 23 , 25 , 29 ].…”

“…In this study, we propose a deep-learning-based EMI noise removal algorithm for use on PA images acquired by a newly constructed PAE system [ 29 ]. Although multiple studies have applied artificial intelligence (AI) techniques to PAT, all of them were related to other topics, such as PA image classification [ 30 , 31 ], reverberation removal [ 32 ], missing data restoration [ 33 , 34 ], artifact removal [ 35 , 36 , 37 ], reconstruction assistance [ 38 , 39 ], image segmentation [ 40 , 41 , 42 ], and resolution enhancement [ 43 , 44 ]; more details on the previous works in this area are provided in Supplementary Table S1 .…”

“…To remove EMI noise from a PAE image, simple methods, such as cross-correlation [ 23 ] or a transverse signal gradient-based noise detection algorithm [ 29 ], have been applied as self-rescue methods in previous studies. We guess that the first method might have worked satisfactorily because the related endoscopic probe was operated in an acoustic-resolution (AR) PAE mode, in which it is typical that PA signals from capillaries are hardly captured and resolved.…”

“…Consequently, the denoising work required a great deal of time and labor for manual-based image segmentation in order to remove the EMI noise that occurred only in the non-intestine area while avoiding the alteration or loss of important capillary signals. Thus, it is our point that, if the operation mode of the PAE or IVPA probe approaches such an optical microscopy level of high resolution, related EMI noise removal becomes tricky, regardless of the type of the applied methods [ 23 , 29 ], because the noise pattern itself becomes similar to that of the capillary signal resolved by the OR PAE; note that, among previous PAE or IVPA studies, in which developing a narrow diameter imaging probe was the key, only three reports demonstrated an OR imaging mode [ 21 , 24 , 29 ].…”