Faculty Publications

Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736

Publications by NITK Faculty

Browse

Subject: search.filters.subject.Data augmentation

Search Results

Now showing 1 - 5 of 5
  • No Thumbnail Available
    Item
    Image Manipulation Detection Using Augmentation and Convolutional Neural Networks
    (Springer Science and Business Media Deutschland GmbH, 2024) Maheshwari, A.; Jain, R.; Mahapatra, R.; Palakuru, S.; Anand Kumar, M.A.
    Image tampering is now simpler than ever, thanks to the explosion of digital photos and the creation of easy image modification tools. As a result, if the situation is not handled properly, the major problems may arise. Many computer vision and deep learning strategies have been put out over the years to address the problem. Having said that, people can easily recognize the photographs that were used in that research. This begs the key question of how CNNs might do on more difficult samples. In this chapter, we build a complex CNN network and use various machine learning algorithms to classify the images and compare the accuracies obtained by them. Its performance is also compared on two different datasets. Additionally, we assess the impact of various hyperparameters and a data augmentation strategy on classification performance. This leads to a conclusion that performance can be considerably impacted by dataset difficulty. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
  • No Thumbnail Available
    Item
    Image Augmentation Strategies to Train GANs with Limited Data
    (Springer Science and Business Media Deutschland GmbH, 2023) Lanka, S.; Velingkar, G.; Varadarajan, R.; Anand Kumar, M.
    Training modern generative adversarial networks (GANs) to produce high-quality images requires massive datasets, which are challenging to obtain in many real-world scenarios, like healthcare. Training GANs on a limited dataset overfits the discriminator on the data to the extent that it cannot correctly distinguish between real and fake images. This paper proposes an augmentation mechanism to improve the dataset’s size, quality, and diversity using a set of different augmentations, namely flipping of images, rotations, shear, affine transformations, translations, and a combination of these to form some hybrid augmentation. Fretchet distance has been used as the evaluation metric to analyze the performance of different augmentations on the dataset. It is observed that as the number of augmentations increase, the quality of generated images improves, and the Fretchet distance reduces. The proposed augmentations successfully improve the quality of generated images by the GAN when trained with limited data. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
  • No Thumbnail Available
    Item
    Classification of Skin Cancer Images using Lightweight Convolutional Neural Network
    (Institute of Electrical and Electronics Engineers Inc., 2023) Sandeep Kumar, T.; Annappa, B.; Dodia, S.
    Skin is the most powerful shield human organ that protects the internal organs of the human body from external attacks. This important organ is attacked by a diverse range of microbes such as viruses, fungi, and bacteria causing a lot of damage to the skin. Apart from these microbes, even dust plays important role in damaging skin. Every year several people in the world are suffering from skin diseases. These skin diseases are contagious and spread very fast. There are varieties of skin diseases. Thus it requires a lot of practice to distinguish the skin disease by the doctor and provide treatment. In order to automate this process several deep learning models are used in recent past years. This paper demonstrates an efficient and lightweight modified SqueezeNet deep learning model on the HAM10000 dataset for skin cancer classification. This model has outperformed state-of-the-art models with fewer parameters. As compared to existing deep learning models, this SqueezeNet variant has achieved 99.7%, 97.7%, and 97.04% as train, validation, and test accuracies respectively using only 0.13 million parameters. © 2023 IEEE.
  • No Thumbnail Available
    Item
    Enhancing Speech De-Identification with LLM-Based Data Augmentation
    (Institute of Electrical and Electronics Engineers Inc., 2024) Dhingra, P.; Agrawal, S.; Veerappan, C.S.; Chng, E.S.; Tong, R.
    This paper addresses the challenge of data scarcity in speech de-identification by introducing a novel, fully automated data augmentation method leveraging large language models. Our approach overcomes the limitations of human annotation, enabling the creation of extensive training datasets. To enhance de-identification performance, we compare pipeline and end-to-end models. While the pipeline approach sequentially applies speech recognition and named entity recognition, the end-to-end model jointly learns these tasks. Experimental results demonstrate the effectiveness of our data augmentation strategy and the superiority of the end-to-end model in improving PII detection accuracy and robustness. © 2024 IEEE.
  • No Thumbnail Available
    Item
    Crossover based technique for data augmentation
    (Elsevier Ireland Ltd, 2022) Raj, R.; Mathew, J.; Kannath, S.K.; Rajan, J.
    Background and Objective: Medical image classification problems are frequently constrained by the availability of datasets. “Data augmentation” has come as a data enhancement and data enrichment solution to the challenge of limited data. Traditionally data augmentation techniques are based on linear and label preserving transformations; however, recent works have demonstrated that even non-linear, non-label preserving techniques can be unexpectedly effective. This paper proposes a non-linear data augmentation technique for the medical domain and explores its results. Methods: This paper introduces “Crossover technique”, a new data augmentation technique for Convolutional Neural Networks in Medical Image Classification problems. Our technique synthesizes a pair of samples by applying two-point crossover on the already available training dataset. By this technique, we create N new samples from N training samples. The proposed crossover based data augmentation technique, although non-label preserving, has performed significantly better in terms of increased accuracy and reduced loss for all the tested datasets over varied architectures. Results: The proposed method was tested on three publicly available medical datasets with various network architectures. For the mini-MIAS database of mammograms, our method improved the accuracy by 1.47%, achieving 80.15% using VGG-16 architecture. Our method works fine for both gray-scale as well as RGB images, as on the PH2 database for Skin Cancer, it improved the accuracy by 3.57%, achieving 85.71% using VGG-19 architecture. In addition, our technique improved accuracy on the brain tumor dataset by 0.40%, achieving 97.97% using VGG-16 architecture. Conclusion: The proposed novel crossover technique for training the Convolutional Neural Network (CNN) is painless to implement by applying two-point crossover on two images to form new images. The method would go a long way in tackling the challenges of limited datasets and problems of class imbalances in medical image analysis. Our code is available at https://github.com/rishiraj-cs/Crossover-augmentation © 2022