2021
DOI: 10.1007/s00455-021-10309-0
|View full text |Cite
|
Sign up to set email alerts
|

Detecting Aspiration During FEES with Narrow Band Imaging in a Clinical Setting

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2

Citation Types

0
4
0

Year Published

2022
2022
2023
2023

Publication Types

Select...
3
1
1

Relationship

0
5

Authors

Journals

citations
Cited by 5 publications
(4 citation statements)
references
References 44 publications
0
4
0
Order By: Relevance
“…This would enable continuous training or planned re-training in certain intervals to enhance the model performance, while at the same time reducing annotation effort when compared with that of an undirected approach. Moreover, when gaining research partners who are in possession of a reasonable amount of narrow-band imaging videos showing aspirations [ 23 , 24 ], this could also be used to further facilitate the AI-based detection. Additionally, we could implement other XAI concepts, such as a combination of frame-wise classification for aspiration detection and XAI methods such as GradCAM or Saliency maps [ 3 ].…”
Section: Discussionmentioning
confidence: 99%
See 1 more Smart Citation
“…This would enable continuous training or planned re-training in certain intervals to enhance the model performance, while at the same time reducing annotation effort when compared with that of an undirected approach. Moreover, when gaining research partners who are in possession of a reasonable amount of narrow-band imaging videos showing aspirations [ 23 , 24 ], this could also be used to further facilitate the AI-based detection. Additionally, we could implement other XAI concepts, such as a combination of frame-wise classification for aspiration detection and XAI methods such as GradCAM or Saliency maps [ 3 ].…”
Section: Discussionmentioning
confidence: 99%
“…One recent approach that addresses the problem of human misses for penetration and aspiration detection is narrow-band imaging. It is implemented in certain types of endoscopes, can be used to sharpen the optical contrasts, and has proven to increase the IRR [ 23 , 24 ]. However, all of the described areas for improvement could be addressed when the administration of FEES is combined with the help of an Artificial Intelligence (AI) tool that is capable of giving reproducible and quantitative output based on a frame-by-frame analysis without concentration errors.…”
Section: Introductionmentioning
confidence: 99%
“…This would enable continuous training or planned re-trainings in certain intervals to enhance the model performance, while at the same time reducing annotation effort when compared with an undirected approach. Moreover, when gaining research partners that are in possession of a reasonable amount of narrow band imaging videos showing aspirations [23,24] this could also be used to further facilitate the AI based detection. Additionally, we could also implement other XAI concepts like a combination of frame-wise classification for aspiration detection and XAI methods like GradCAM or Saliency maps [3].…”
Section: Discussionmentioning
confidence: 99%
“…One recent approach that addresses the problem of human misses for penetration and aspiration detection is narrow band imaging. It is implemented in certain types of endoscopes, can be used to sharpen the optical contrasts, and has proven to increase the IRR [23,24]. However, all described aspects for improvement could be addressed, when combining the administration of FEES with the help of an Artificial Intelligence (AI) tool that is capable of giving reproducible, quantitative output based on a frame-by-frame analysis without concentration errors.…”
Section: Introductionmentioning
confidence: 99%