Image copy detection is commonly addressed using either local descriptors or deep learning models, which can be computationally expensive and rely on high-dimensional features. In contrast, this work explores copy detection using compact perceptual hash representations and learned similarity functions defined directly on hash codes. We evaluate classical hash distances under realistic transformations using the PIHD dataset and assess generalization on a modified MS COCO dataset (mCOCO). We propose SiPHaD, a Siamese-based model that learns similarity in hash space, improving retrieval performance while maintaining efficiency. Results demonstrate that lightweight hash-based approaches, when combined with learned similarity, provide a strong alternative to feature-heavy pipelines.
Abstract: Brain tumors are some of the most serious issues affecting the central nervous system. They need to be detected early and correctly. Traditional machine learning techniques rely on collecting data in one place, which raises significant privacy and security concerns. Thi…
Background and Objective: Self-harm is a psychologically damaging behavior, and its accurate differentiation from other wounds (violence, accidents, burns, diabetic ulcers) is critically important in forensic medicine. However, this differentiation often falls into a diagnostic "…
Self-supervised deep learning has emerged as a powerful method for image enhancement when a priori ground-truth references are not available. Stemming from Noise2Noise , it was shown that a convolutional neural network (CNN) can be trained from a noisy input and target pair of th…
Abstract: Video surveillance has become a critical component in today's world. With developments in advanced systems have been developed as a result of the precision and efficacy of deep learning, machine learning, and artificial intelligence to identify and identify questionable…
Large language models (LLMs) are increasingly deployed in user-facing applications, which exposes them to prompt injection and jailbreak attacks that override system instructions, exfiltrate data, or elicit disallowed behaviour. Existing defences are largely single-mechanism: a r…
Real-world phishing traffic is heavily imbalanced — legitimate URLs vastly outnumber phishing ones in any live traffic stream — and the two error types carry different costs: a missed phishing site (false negative) can lead directly to credential theft, while a legitimate site wr…