Scholarlink

In this paper, we present a comprehensive study on deepfake detection techniques to summarize and identify the best-performing state-of-the-art methods. We try to dig deep into the solutions out there at present and delve into what are they good for? We also present a new architecture specifically designed for deepfake detection pipelines, with novel features to increase accuracy and robustness. Experimental results also show the effectiveness of our method by comparing it with traditional methods. Our results provide insights into the recent trends, limitations, and future directions for deepfake detection research. This paper could be referred to as a guide for researchers, practitioners, and policymakers who are fighting against the proliferation of synthetic media manipulations where it helps synthesize insights from various methodologies and empirical studies. Moreover, we introduced an example hybrid deepfake detection model using EfficientNetB4 for feature extraction and Long Short-Term Memory (LSTM) for classification.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Novel Hybrid Deep Learning Technique for Deepfake Detection: A Review Analysis and Proposed Approach

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Shital S. Bhandare,
  • Kamini A. Shirsath

摘要

In this paper, we present a comprehensive study on deepfake detection techniques to summarize and identify the best-performing state-of-the-art methods. We try to dig deep into the solutions out there at present and delve into what are they good for? We also present a new architecture specifically designed for deepfake detection pipelines, with novel features to increase accuracy and robustness. Experimental results also show the effectiveness of our method by comparing it with traditional methods. Our results provide insights into the recent trends, limitations, and future directions for deepfake detection research. This paper could be referred to as a guide for researchers, practitioners, and policymakers who are fighting against the proliferation of synthetic media manipulations where it helps synthesize insights from various methodologies and empirical studies. Moreover, we introduced an example hybrid deepfake detection model using EfficientNetB4 for feature extraction and Long Short-Term Memory (LSTM) for classification.