This chapter investigates image-based malware family classification using an ensemble learning framework and a soft voting strategy, and finds that the best-performing soft voting pool--consisting of fifteen voters selected on a dedicated validation split--achieves 80.2% accuracy across the 17 malware families under consideration.
Abstract
In this chapter, we investigate image-based malware family classification using an ensemble learning framework and a soft voting strategy. We consider malware binaries that have been converted into images using eight distinct conversion strategies. Three complementary feature extraction tracks are applied to these images: handcrafted descriptors combining Histogram of Oriented Gradients (HOG) and Haralick texture features along with 38 statistical features; dense embeddings obtained from three pretrained neural networks (VGG16, ResNet50, and ViT-B/16), where each pretrained model is used as a frozen feature extractor with its classification head removed; and 512-dimensional embeddings derived from a custom Convolutional Neural Network (CNN) trained directly on the malware images. Each of the three feature extraction techniques is evaluated with machine learning classifiers across all eight image conversion types. The best individual results are 77.8% accuracy for the handcrafted features, 73.8% for the pretrained neural network track, and 74.8% for the custom CNN track. Then we consider various soft voting ensemble strategies, and we find that the best-performing soft voting pool--consisting of fifteen voters selected on a dedicated validation split--achieves 80.2% accuracy across the 17 malware families under consideration, a statistically significant improvement of 2.4 percentage points over the best individual model. A quantitative diversity analysis confirms that the different feature representations are complementary, with the handcrafted descriptors being the strongest contributors.
Image-based analysis has been extensively used for malware classification
by converting binary files into grayscale images to extract structural and behavioral patterns.
However, common deep learning approaches generally suffer from problems such as
non-interpretability, class imbalance, and computational overhead....
Shajjia Mazhar, Sadia Fatima, Irfanud Din et al.· Recent Advances in Computer...· 0 citations
The high rate of growth of advanced and highly obfuscated malware has made the use of conventional signature detection mechanisms less viable. In order to overcome this problem, this paper suggests a hybrid framework for malware classification based on the combination of Convolutional Neural Networks (CNNs) and Long Sh...
Chevella Anil Kumar, Saradha S, Mithila Ayyavoo et al.· International Research Journ...· 0 citations
A novel hybrid CNN-Hopfield Neural Network (CNN-HNN) framework that replaces traditional dense classification heads with continuous Modern Hopfield associative memory layers is proposed, achieving top-tier accuracy and reducing classification-head parameter counts while reducing floating-point operations (FLOPs) and ac...
Uma Kannan, Rajendran Swamidurai· International journal of res...· 0 citations
Early and correct detection of cancer regions in histopathology images is important for timely diagnosis and treatment. In this work, we proposed a Explainable Artificial Intelligence based convolutional Neural network (CNN) framework for the classification of lung and colon cancer. The framework combines deep features...
Onkar Singh, Chandrasen Pandey, K. Singh et al.· Discover Artificial Intellig...· 0 citations
The increasing sophistication of modern malware, particularly in the form of polymorphic and packed variants, poses significant challenges to traditional detection systems. While machine learning-based approaches have shown promise, many existing methods rely on single feature representations and are evaluated using me...
Akshaya Mogili, G. Lishita, R. N· 2026 International Conferenc...· 0 citations
The findings show that the hybrid CNN–LVQ approach can be applied to chili seed classification without requiring specialized hardware or a pre-trained model.