國立虎尾科技大學 |

Enhancing Robotic Vision Through Excitatory-Inhibitory Recurrent Memory Networks : = Facial Recognition and Recall.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Enhancing Robotic Vision Through Excitatory-Inhibitory Recurrent Memory Networks :/
Reminder of title:	Facial Recognition and Recall.
Author:	Srivastava, Atin.
Description:	1 online resource (59 pages)
Notes:	Source: Masters Abstracts International, Volume: 85-11.
Contained By:	Masters Abstracts International85-11.
Subject:	Computer science. -
Online resource:	click for full text (PQDT)
ISBN:	9798382758343

Enhancing Robotic Vision Through Excitatory-Inhibitory Recurrent Memory Networks : = Facial Recognition and Recall.
Srivastava, Atin.

Enhancing Robotic Vision Through Excitatory-Inhibitory Recurrent Memory Networks :Facial Recognition and Recall. - 1 online resource (59 pages)

Source: Masters Abstracts International, Volume: 85-11.

Thesis (M.S.)--Rutgers The State University of New Jersey, School of Graduate Studies, 2024.

Includes bibliographical references

Robotic vision, characterized by the necessity for real-time processing of complex, dynamic data in ever-changing environments, confronts significant challenges with traditional computational paradigms. Deep learning algorithms, while powerful, often rely on static conditions and extensive, homogenous datasets, limitations that are impractical in real-world scenarios. Addressing these challenges, previous work introduces a neuromorphic oculomotor controller integrated into a biomimetic robotic head prototype, inspired by neurological studies that elucidate the unified circuitry behind saccade and smooth pursuit eye movements. In this work, we present an innovative approach that leverages excitatory-inhibitory recurrent memory networks for facial recognition and recall, enhancing robotic vision beyond conventional methodologies. Building upon this foundation, our study extends the capabilities of neuromorphic systems through the development and integration of an Excitatory-Inhibitory Recurrent Memory Networks (EIR-MN) model. This model further enhances the biomimetic aspects of our robotic head prototype by introducing a mechanism for facial recognition, even in noisy environments. The EIR-MN model showcases resilience to environmental noise through its dynamic temporal processing capabilities. Its architecture, designed to reflect the intricacies of neural processing found in the human brain, leverages spike-based encoding to achieve high levels of accuracy in facial recognition tasks. This model not only complements the previously developed oculomotor controller by adding visual processing abilities but also underscores the potential of neuromorphic computing to emulate complex cognitive functions without the need for extensive training datasets. The controller is unique in its utilization of spikes for encoding and processing information among models of biological neurons, and its architecture mimics brain anatomy, enabling operation without specific training. The EIR-MN model will be running on top of this previous work. This advancement aligns with ComBra Lab's vision of developing neuro-mimetic "bottom-up" computational models of brain networks that harness natural intelligence for robotic control. Our contributions extend the robustness of the robot in target detection, particularly in facial recognition tasks, broadening the spectrum of applications and marking a significant step forward in the pursuit of biologically inspired, efficient robotic vision systems. This work not only enhances robotic capabilities in complex environments but also offers insights into the potential convergence of neuroscience, robotics, and artificial intelligence.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798382758343Subjects--Topical Terms:

573171
Computer science.
Subjects--Index Terms:

Computational paradigmsIndex Terms--Genre/Form:

554714
Electronic books.

Enhancing Robotic Vision Through Excitatory-Inhibitory Recurrent Memory Networks : = Facial Recognition and Recall.
LDR:04159ntm a22003977 4500 001 1149708
005 20241022112620.5
006 m o d
007 cr bn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798382758343
035 $a (MiAaPQ)AAI31143313
035 $a AAI31143313
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Srivastava, Atin. $3 1476039
245 1 0 $a Enhancing Robotic Vision Through Excitatory-Inhibitory Recurrent Memory Networks : $b Facial Recognition and Recall.
264 0 $c 2024
300 $a 1 online resource (59 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Masters Abstracts International, Volume: 85-11.
500 $a Advisor: Michmizos, Konstantinos.
502 $a Thesis (M.S.)--Rutgers The State University of New Jersey, School of Graduate Studies, 2024.
504 $a Includes bibliographical references
520 $a Robotic vision, characterized by the necessity for real-time processing of complex, dynamic data in ever-changing environments, confronts significant challenges with traditional computational paradigms. Deep learning algorithms, while powerful, often rely on static conditions and extensive, homogenous datasets, limitations that are impractical in real-world scenarios. Addressing these challenges, previous work introduces a neuromorphic oculomotor controller integrated into a biomimetic robotic head prototype, inspired by neurological studies that elucidate the unified circuitry behind saccade and smooth pursuit eye movements. In this work, we present an innovative approach that leverages excitatory-inhibitory recurrent memory networks for facial recognition and recall, enhancing robotic vision beyond conventional methodologies. Building upon this foundation, our study extends the capabilities of neuromorphic systems through the development and integration of an Excitatory-Inhibitory Recurrent Memory Networks (EIR-MN) model. This model further enhances the biomimetic aspects of our robotic head prototype by introducing a mechanism for facial recognition, even in noisy environments. The EIR-MN model showcases resilience to environmental noise through its dynamic temporal processing capabilities. Its architecture, designed to reflect the intricacies of neural processing found in the human brain, leverages spike-based encoding to achieve high levels of accuracy in facial recognition tasks. This model not only complements the previously developed oculomotor controller by adding visual processing abilities but also underscores the potential of neuromorphic computing to emulate complex cognitive functions without the need for extensive training datasets. The controller is unique in its utilization of spikes for encoding and processing information among models of biological neurons, and its architecture mimics brain anatomy, enabling operation without specific training. The EIR-MN model will be running on top of this previous work. This advancement aligns with ComBra Lab's vision of developing neuro-mimetic "bottom-up" computational models of brain networks that harness natural intelligence for robotic control. Our contributions extend the robustness of the robot in target detection, particularly in facial recognition tasks, broadening the spectrum of applications and marking a significant step forward in the pursuit of biologically inspired, efficient robotic vision systems. This work not only enhances robotic capabilities in complex environments but also offers insights into the potential convergence of neuroscience, robotics, and artificial intelligence.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Computer science. $3 573171
650 4 $a Robotics. $3 561941
653 $a Computational paradigms
653 $a Homogenous datasets
653 $a Deep learning
653 $a Neuromorphic oculomotor
653 $a Robotic vision systems
655 7 $a Electronic books. $2 local $3 554714
690 $a 0800
690 $a 0984
690 $a 0771
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Rutgers The State University of New Jersey, School of Graduate Studies. $b Computer Science. $3 1241263
773 0 $t Masters Abstracts International $g 85-11.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31143313 $z click for full text (PQDT)