Conference Papers

Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506

Browse

Subject: search.filters.subject.Convolutional Neural Networks (CNNs)

Search Results

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Exploring Convolutional Neural Networks for Image Classification and Object Detection
    (Institute of Electrical and Electronics Engineers Inc., 2024) Sadhankar, D.S.; Illa, M.; Shetty, P.; Kumar, S.V.; Megha, M.K.; Ambilwade, R.P.
    Convolutional Neural Networks or CNNs are one of the newest powerful tools in various tasks of computer vision such as image classification or object detection providing the highest accuracy. This paper also aims to evaluate the efficiency of the CNNs in these areas using a real-world dataset from Kaggle. We discuss general issues and ways to address it, such as data augmentation, dropout and choose the best/settled value of hyperparameters for improvement of the model. This paper aimed at analyzing the effectiveness of CNN in learning discriminative features from the images and confirm that CNNs are among the most accurate models for image classification. Moreover, this study also provides suggestions for subsequent research work, including improving the CNN architectures, employing transfer learning, and incorporating interpretation methods to continue enhancing the performance of CNNs in computer vision. © 2024 IEEE.
  • No Thumbnail Available
    Item
    Addressing Diffusion Model Based Counter-Forensic Image Manipulation for Synthetic Image Detection
    (Association for Computing Machinery, 2024) Herur, A.N.; Santhosh, V.; Shetty, N.; Seelamantula, C.S.
    With the rapid development of modern generative models, the need for an automated synthetic image detection process has never been greater. Recent works in the field of synthetic image detection focus on improving out-of-distribution (OoD) classification performance and robustness to common image pre-processing techniques. However, in this work, we intend to explore the nature of an intricate counter-forensic attack, i.e., the reconstruction of real images with Diffusion Model autoencoders, which could be used to adversely affect the performance of modern synthetic image detection algorithms. We present a variety of experiments to study the nature of this counter-forensic attack and use the inferences from these experiments to develop multiple algorithms to detect such reconstructed images while attempting to detect real and purely synthetic images accurately. To do so, we make use of trained classifiers that can detect real images, autoencoder-reconstructed images, and purely synthetic images. Furthermore, we combine these techniques to build a novel ensemble algorithm that competes with state-of-the-art (SoTA) algorithms in the ‘Real vs. Fake’ image detection task, while detecting autoencoder reconstructed images accurately, attaining an accuracy of 99.2% in the multiclass setting. © 2024 Copyright held by the owner/author(s).