Image Hashing Research Articles

Distance distributions and runtime analysis of perceptual hashing algorithms

Perceptual image hashing refers to a class of algorithms that produce content-based image hashes. These systems use specialized perceptual hash algorithms like Phash, Microsoft’s PhotoDNA, or Facebook’s PDQ to generate a compact digest of an image file that can be roughly compared to a database of known illicit-content digests. Time taken by perceptual hashing algorithms to generate hash code has been computed. Then, we evaluated perceptual hashing algorithms on two million dataset of images. The produced nine variants of the original images were computed and then several distances were calculated. There have been several studies in the past, but in the existing literature size of the data is small and there are very few studies with hash code computation time and robustness tradeoff. This work shows that existing perceptual hashing algorithms are robust for most of the content-preserving operations and there is a tradeoff between computation time and robustness.

A robust medical image zero-watermarking algorithm using Collatz and Fresnelet Transforms

Zero-watermarking in medical images is an emerging field that focuses on calculating the invisible data (key) using medical imagery to ensure data integrity and authenticity without compromising diagnostic accuracy. This paper introduces a robust zero-watermarking technique leveraging the Collatz and Fresnelet Transforms. The Forward Collatz Transform (FCT) is initially applied to create a secure and encrypted embedding pattern for medical images. Subsequently, the Fresnelet Transform (FT) is employed, offering superior localization and frequency selectivity. From the fresnelet values, we extract two strongest Oriented FAST and Rotated BRIEF (ORB) points to enhance watermark robustness, resulting in a 64-bit perceptual image hash. Our approach adopts a dual-layer security strategy by combining FCT and Cyclic-Shift-Transformation (CST) methods, significantly fortifying the protection of watermark image data. The watermark can be efficiently extracted using the Inverse Collatz Transform (ICT). A comprehensive performance analysis evaluates our system under single, double, and multiple attacks on medical images. Our experiments clearly show that our system outperforms existing methods in medical image watermarking, demonstrating its resilience against various manipulations. This approach can significantly improve data security and reliability in medical imaging applications.

A New Distance Measurement Method for Perceptual Image Hashing

A large number of perceptual image hashing schemes have been developed in recent decades, and they usually calculated the hash distances using Euclidean distance, correlation coefficient, normalized Hamming distance, etc. These hash distances are not customized measurements for hash sequences, which are all single values. To this end, a new distance measurement method, i.e. a distance vector instead of a single value, is proposed in this paper. Since a hash sequence can be regarded as the compressed representation of image perceptual content, the correlation between images is maintained between hash sequences. In our method, the correlation between hash sequences is captured by residuals and co-occurrence matrix to form a distance vector. The correlation differences between perceptually similar and distinct images can be captured in the obtained distance vectors. Using a binary classifier, it is easy to classify the distance vectors of perceptually similar image pairs and perceptually distinct image pairs. As a result, content authentication accuracy of perceptual image hashing can be effectively increased, which is verified by experimental results.

A robust self-supervised image hashing method for content identification with forensic detection of content-preserving manipulations

Image content identification systems have many applications in industry and academia. In particular, a hash-based content identification system uses a robust image hashing function that computes a short binary identifier summarizing the perceptual content in a picture and is invariant against a set of expected manipulations while being capable of differentiating between different pictures. A common approach to designing these algorithms is crafting a processing pipeline by hand. Unfortunately, once the context changes, the researcher may need to define a new function to adapt. A deep hashing approach exploits the feature learning capabilities in deep networks to generate a feature vector that summarizes the perceptual content in the image, achieving outstanding performance for the image retrieval task, which requires measuring semantic and perceptual similarity between items. However, its application to robust content identification systems is an open area of opportunity. Also, image hashing functions are valuable tools for image authentication. However, to our knowledge, its application to content-preserving manipulation detection for image forensics tasks is still an open research area. In this work, we propose a deep hashing method exploiting the metric learning capabilities in contrastive self-supervised learning with a new modular loss function for robust image hashing. Moreover, we propose a novel approach for content-preserving manipulation detection for image forensics through a sensitivity component in our loss function. We validate our method through extensive experimentation in different data sets and configurations, validating the generalization properties in our work.

Deep Supervised Hashing by Fusing Multiscale Deep Features for Image Retrieval

Deep network-based hashing has gained significant popularity in recent years, particularly in the field of image retrieval. However, most existing methods only focus on extracting semantic information from the final layer, disregarding valuable structural information that contains important semantic details, which are crucial for effective hash learning. On the one hand, structural information is important for capturing the spatial relationships between objects in an image. On the other hand, image retrieval tasks often require a more holistic representation of the image, which can be achieved by focusing on the semantic content. The trade-off between structural information and image retrieval accuracy in the context of image hashing and retrieval is a crucial consideration. Balancing these aspects is essential to ensure both accurate retrieval results and meaningful representation of the underlying image structure. To address this limitation and improve image retrieval accuracy, we propose a novel deep hashing method called Deep Supervised Hashing by Fusing Multiscale Deep Features (DSHFMDF). Our approach involves extracting multiscale features from multiple convolutional layers and fusing them to generate more robust representations for efficient image retrieval. The experimental results demonstrate that our method surpasses the performance of state-of-the-art hashing techniques, with absolute increases of 11.1% and 8.3% in Mean Average Precision (MAP) on the CIFAR-10 and NUS-WIDE datasets, respectively.

Image Hash Layer Triggered CNN Framework for Wafer Map Failure Pattern Retrieval and Classification

Recently, deep learning methods are often used in wafer map failure pattern classification. CNN requires less feature engineering but still needs preprocessing, e.g., denoising and resizing. Denoising is used to improve the quality of the input data, and resizing is used to transform the input into an identical size when the input data sizes are various. However, denoising and resizing may distort the original data information. Nevertheless, CNN-based applications are focusing on studying different feature map architectures and the input data manipulation is less attractive. In this study, we proposed an image hash layer triggered CNN framework for wafer map failure pattern retrieval and classification. The motivation and novelty are to design a CNN layer that can play as a resizing, information retrieval-preservation method in one step. The experiments proved that the proposed hash layer can retrieve the failure pattern information while maintaining the classification performance even though the input data size is decreased significantly. In the meantime, it can prevent overfitting, false negatives, and false positives, and save computing costs to a certain extent.

Robust Image Hashing in Encrypted Domain

Robust Image Hashing in Encrypted Domain

Prediction of Flipped Bits in Robust Image Hashes by Machine Learning

Prediction of Flipped Bits in Robust Image Hashes by Machine Learning

Pyramid hybrid pooling quantization for efficient fine-grained image retrieval

Deep hashing approaches, including deep quantization and deep binary hashing, have become a common solution to large-scale image retrieval due to their high computation and storage efficiency. Most existing hashing methods cannot produce satisfactory results for fine-grained retrieval, because they usually adopt the outputs of the last CNN layer to generate binary codes. Since deeper layers tend to summarize visual clues, e.g., texture, into abstract semantics, e.g., dogs and cats, the feature produced by the last CNN layer is less effective in capturing subtle but discriminative visual details that mostly exist in shallow layers. To improve fine-grained image hashing, we propose Pyramid Hybrid Pooling Quantization (PHPQ). Specifically, we propose a Pyramid Hybrid Pooling (PHP) module to capture and preserve fine-grained semantic information from multi-level features, which emphasizes the subtle discrimination of different sub-categories. Besides, we propose a learnable quantization module with a partial codebook attention mechanism, which helps to optimize the most relevant codewords and improves the quantization. Comprehensive experiments on two widely-used public benchmarks, i.e., CUB-200-2011 and Stanford Dogs, demonstrate that PHPQ outperforms state-of-the-art methods.

SDTU-Net: Stepwise-Drop and Transformer-Based U-Net for Subject-Sensitive Hashing of HRRS Images

SDTU-Net: Stepwise-Drop and Transformer-Based U-Net for Subject-Sensitive Hashing of HRRS Images

Exploring Hierarchical Information in Hyperbolic Space for Self-Supervised Image Hashing.

In real-world datasets, visually related images often form clusters, and these clusters can be further grouped into larger categories with more general semantics. These inherent hierarchical structures can help capture the underlying distribution of data, making it easier to learn robust hash codes that lead to better retrieval performance. However, existing methods fail to make use of this hierarchical information, which in turn prevents the accurate preservation of relationships between data points in the learned hash codes, resulting in suboptimal performance. In this paper, our focus is on applying visual hierarchical information to self-supervised hash learning and addressing three key challenges, including the construction, embedding, and exploitation of visual hierarchies. We propose a new self-supervised hashing method named Hierarchical Hyperbolic Contrastive Hashing (HHCH), making breakthroughs in three aspects. First, we propose to embed continuous hash codes into hyperbolic space for accurate semantic expression since embedding hierarchies in the hyperbolic space generates less distortion than in the hyper-sphere or Euclidean space. Second, we update the K-Means algorithm to make it run in the hyperbolic space. The proposed hierarchical hyperbolic K-Means algorithm can achieve the adaptive construction of hierarchical semantic structures. Last but not least, to exploit the hierarchical semantic structures in hyperbolic space, we propose the hierarchical contrastive learning algorithm, including hierarchical instance-wise and hierarchical prototype-wise contrastive learning. Extensive experiments on four benchmark datasets demonstrate that the proposed method outperforms state-of-the-art self-supervised hashing methods. Our codes are released at https://github.com/HUST-IDSM-AI/HHCH.git.

Two Robust Perceptual Image Hashing Schemes Based on Discrete Wavelet Transform

Two Robust Perceptual Image Hashing Schemes Based on Discrete Wavelet Transform

High-Performance Encryption Algorithms for Dynamic Images Transmission

With the proliferation of the internet, the issue of image tampering has escalated, necessitating robust image encryption schemes. Despite the multitude of proposed image encryption algorithms, challenges such as slow computational speed, weak security, and suboptimal visibility persist. This study addresses these challenges by introducing a high-performance encryption algorithm tailored for dynamic images—QEDI (Quick Encryption Algorithm for Dynamic Images). QEDI leverages elliptic curve-based asymmetric encryption algorithms and hash algorithms. The process involves using a hash algorithm to calculate the hash of the dynamic image slated for encryption, employing the elliptic curve algorithm to generate public and private keys, and utilizing the asymmetric encryption algorithm to process the image hash, resulting in a signature ciphertext. This ciphertext is embedded into the dynamic image, completing the encryption process. To enhance the execution speed without compromising the image quality and visibility, the bit depth of the dynamic image is reduced before hash calculation, and custom information fields are encapsulated to embed the signature ciphertext into the dynamic image. Experimental evaluations, conducted within a dynamic image generation system, encompassed assessments of security, encryption algorithm execution time, and image quality. The results indicate that utilizing QEDI for dynamic image encryption yields better security compared to EDI (Encryption Algorithm for Dynamic Images). QEDI exhibits minimal impact on image quality, with a noteworthy 69.77% reduction in execution time compared to EDI. Furthermore, when compared to existing algorithms employing image conversion for encryption, QEDI-encrypted images demonstrate enhanced visibility, better image quality, and expedited encryption speed.

A lightweight robust image hash based on random tensors and angle features for IoT devices

A lightweight robust image hash based on random tensors and angle features for IoT devices

Deep Learning-Based Image Retrieval With Unsupervised Double Bit Hashing

Unsupervised image hashing is a widely used technique for large-scale image retrieval. This technique maps an image to a finite length of binary codes without extensive human-annotated data for compact storage and effective semantic retrieval. This study proposes a novel deep unsupervised double-bit hashing method for image retrieval. This approach is based on the double-bit hashing method, which has been shown to better preserve the neighboring structure of binary codes than single-bit hashing. Traditional double-bit hashing methods require the entire dataset to be processed simultaneously to determine optimal thresholding values of binary feature encoding. In contrast, the proposed method trains the hashing layer in a minibatch manner, allowing for adaptive threshold learning through a gradient-based optimization strategy. Additionally, unlike most former methods, which only train the hashing networks on top of fixed pre-trained neural networks backbone. The proposed learning framework trains both hashing and backbone networks alternately asynchronously. This strategy enables the model to maximize the learning capability of the hashing and backbone networks. Furthermore, adopting the lightweight Vision Transformer (ViT) in the proposed method allows the model to capture both local and global relationships between multiple image views exemplar, which lead to better generalization, thus maximizing the retrieval performance of the model. Extensive experiments on CIFAR10, NUW-WIDE, and FLICKR25K datasets validate that the proposed method has superior retrieval quality and computational efficiency than state-of-the-art methods.

PIH-mSCM: a modified spiking cortical model for perceptual image hashing and its application to copy detection

PIH-mSCM: a modified spiking cortical model for perceptual image hashing and its application to copy detection

Comparison of deep learning classification models for facial image age estimation in digital forensic investigations

There has been a significant rise in digital forensic investigations containing Indecent Images of Children (IIoC), and one of the major challenges faced by investigators is the time-consuming task of manually investigating images for illicit content. In the UK, law enforcement maintains and uses a standard national repository of IIoC, known as CAID (Child Abuse Image Database), to identify known illegal images by matching their image hashes and metadata. The CAID plays a significant role in making IIoC investigations faster and more effective. However, all images that are not matched through using CAID require manual analysis. Every image has to be viewed and verified as IIoC by investigators. The victim age estimation in the images (i.e., determining whether they are juvenile or adult as this would change the course of the investigation) is a crucial part of this verification process and takes time due to a large number of images to inspect, therefore impacting the speed of the investigation, and consequently victims. This is a time-consuming and challenging task for human investigators.Previous work has demonstrated that deep learning has the capability to estimate age with high accuracy in images. This reduces the number of images that will need to be manually processed, thereby finishing the investigation faster. However, in terms of practical implementation in IIoC investigations, there is an absence of a comparative study using the same datasets to establish the most appropriate deep learning model and classification approach to use. This is important as different models have different capabilities and previous works utilise various binary, multi-class, and regression approaches. It is not yet known which is the most accurate for use in digital forensic investigations. In this paper, we construct an extensive dataset before experimenting with four pre-trained deep learning models: VGG16, ResNet50, Xception, and InceptionV3. We have identified that binary classification works best for the identification of images as a child or adult, with the ResNet50 obtaining the best results in terms of accuracy (91.70%) on unseen images.

Sequential and Correlated Image Hash Code Generation with Deep Reinforcement Learning

Image hashing is an algorithm used to represent an image with a unique value. Hashing methods, which are generally developed to search for similar examples of an image, have gained a new dimension with the use of deep network structures and better results have started to be obtained with the methods. The developed deep network models generally consider hash functions independently and do not take into account the correlation between them. In addition, most of the existing data-dependent hashing methods use pairwise/triplet similarity metrics that capture data relationships from a local perspective. In this study, the Central similarity metric, which can achieve better results, is adapted to the deep reinforcement learning method with sequential learning strategy, and successful results are obtained in learning binary hash codes. By taking into account the errors of previous hash functions in the deep reinforcement learning strategy, a new model is presented that performs interrelated and central similarity based learning.

Robust image hashing based on multi-view dimension reduction

In the field of multimedia security, image hashing is an effective technology for solving image content identification. Image hashing algorithm aims to map the high dimensional input features onto perceptual hashing values in terms of their perceptual content, in which visually identical images are mapped to similar digital representations. In this paper, we focus on multi-view dimensionality reduction based hashing generation, and propose a novel method, called Multi-view Dimensionality Reduction based Robust Image Hashing (MDRIH). Our MDRIH scheme maps the multi-view features into a compact binary space by considering the complementarity and similarity between image features across these different views. For preprocessing, bilinear interpolation is applied on the host image to adjust to a standard size, and ensure the different images with same hash size. Specifically, in order to improve the robustness of the algorithm against content-preserving operations, multi-view features, including structural features, edge features, and color features are extracted. After that, multi-view dimension reduction is adopted to obtain meaningful low-dimensional information from multi-feature fusion. Finally, the data in low-dimensional subspace is encoded into a compact binary sequence. The experimental results indicate that the proposed hashing method not only preserves the correlation between multiple views and the similarity among different data points within each view, but also achieves a better balance between robustness and discriminability compared to existing hashing methods.

Analysis of Image Hashing in Wafer Map Failure Pattern Recognition

There are mainly two types of wafer map failure pattern recognition, i.e., traditional classification based and deep learning based approaches. Traditional classification usually needs feature engineering, and deep learning requires lower human intervention and feature engineering. The joint requirement is noise filtering and identical input dimension size. Feature engineering and dimension resizing are artificial work and vary from study to study. In our study, we proposed an image hashing based wafer map resizing and pattern information retrieval framework. Traditional classification methods and CNN are employed for the wafer map failure pattern recognition. The experiments show that the image hashing based framework is the more appropriate method when compared to feature extraction and image resizing methods. We also found that proper input data manipulation can result in acceptable performance whether the used method is traditional classification algorithms or CNN.