Leveraging Deep Learning Architectures for Deepfake Audio Analysis DOI Open Access

Siva Kumar C,

Siva Nageswara Rao G,

Vivek Patnam

et al.

Kalpa publications in computing, Journal Year: 2024, Volume and Issue: 19, P. 266 - 254

Published: Aug. 6, 2024

Deepfake content is created or changed artificially utilizing AI strategies to make it genuine. This research addresses the evolving challenge of detecting deepfake audio content, as recent advancements in technology have rendered increasingly challenging distinguish fabricated content. Leveraging machine and deep learning methodologies, specifically employing Mel-frequency cepstral coefficients (MFCCs) for sound component extraction, we focus on Genuine-or-Fake dataset — a cutting-edge benchmark generated through text- to-speech (TTS) model. arranged into sub-datasets because length spot rate. study reveals that Convolutional Neural Network (CNN) models exhibit highest accuracy identifying within for-rerec for-2-sec datasets. Meanwhile, gradient boosting model performs well for-norm dataset. illustrates CNN model's outstanding performance for-original dataset, outperforming other models. advances field recognition, especially areas manipulation, demonstrating efficacy fake

Language: Английский

Robust Anomaly Detection in Network Traffic using Deep Learning Models DOI

Ashish Nema,

Raghvendra Singh Tomar,

Mani Anand

et al.

Published: Dec. 8, 2023

The exact detection of anomalies in computer network traffic is crucial to protection. This study presents a novel approach achieving the stated goal: use deep learning models. To correctly capture temporal, geographic, and probabilistic aspects data, recently developed integrates autoencoders (DAE), variable (VAE), long short-term memory (LSTM) networks. proposed strategy surpassed six industry-standard solutions terms accuracy, recall, F1-score, AUC-ROC, false positive rate (FPR), negative (FNR). was demonstrated via performance reviews. Furthermore, technique makes optimum currently available resources. improves security by employing more robust anomaly algorithms.

Language: Английский

Citations

1
Cloud-Based Anomaly Detection for Broken Rail Track Using LSTM Autoencoders and Cross-modal Audio Analysis DOI
Smita Rath,

H.C. Upadhyay,

Somya Prakash

et al.

Lecture notes in networks and systems, Journal Year: 2024, Volume and Issue: unknown, P. 69 - 84

Published: Jan. 1, 2024

Language: Английский

Citations

0
VidAnomalyNet: An Efficient Anomaly Detection in Public Surveillance Videos Through Deep Learning Architectures DOI Creative Commons

K Chidananda,

A. P. Siva Kumar

International Journal of Safety and Security Engineering, Journal Year: 2024, Volume and Issue: 14(3), P. 953 - 966

Published: June 24, 2024

In the contemporary era, computer vision applications assume significance due to their role in real world.Video surveillance is one such application that has become indispensable with plenty of unprecedented applications.Detection abnormal events from videos time its importance like traffic monitoring, crime investigation, public safety, healthcare and operations management mention few.With emergence Artificial Intelligence (AI) automatic video taken next level sophistication learning detection anomalies.Particularly deep model Convolutional Neural Network (CNN) found more appropriate for image processing.However, as size does not fit all, CNN provide acceptable accuracy unless it enhanced suitable number layers configurations.Towards this end, paper, we proposed a novel architecture known VidAnomalyNet which based on model.It designed have process anomalies videos.We framework exploit our leveraging performance.We also an algorithm Automatic Anomaly Detection (VAAD).Automatic anomaly context networks refers use computational methods automatically identify unusual or patterns within sequence frames.The goal develop models can distinguish between normal activities unexpected anomalies.Video crucial various applications, including surveillance, industrial safety.At present, detects three classes fire, accident robbery.It be easily extended anomalies.We explored MobileNetV1 transfer by adding new base detection.Our empirical study revealed outperforms MobileNetV1.Highest achieved 96.35%.

Language: Английский

Citations

0
Home Security and Anomaly Detection System: A Comprehensive Solution DOI

S Sophia,

A Princely Nesaraj,

Arpit Raj

et al.

Published: April 18, 2024

Language: Английский

Citations

0
Leveraging Deep Learning Architectures for Deepfake Audio Analysis DOI Open Access

Siva Kumar C,

Siva Nageswara Rao G,

Vivek Patnam

et al.

Kalpa publications in computing, Journal Year: 2024, Volume and Issue: 19, P. 266 - 254

Published: Aug. 6, 2024

Deepfake content is created or changed artificially utilizing AI strategies to make it genuine. This research addresses the evolving challenge of detecting deepfake audio content, as recent advancements in technology have rendered increasingly challenging distinguish fabricated content. Leveraging machine and deep learning methodologies, specifically employing Mel-frequency cepstral coefficients (MFCCs) for sound component extraction, we focus on Genuine-or-Fake dataset — a cutting-edge benchmark generated through text- to-speech (TTS) model. arranged into sub-datasets because length spot rate. study reveals that Convolutional Neural Network (CNN) models exhibit highest accuracy identifying within for-rerec for-2-sec datasets. Meanwhile, gradient boosting model performs well for-norm dataset. illustrates CNN model's outstanding performance for-original dataset, outperforming other models. advances field recognition, especially areas manipulation, demonstrating efficacy fake

Language: Английский

Citations

0