Image Spam Research Articles

Images Spam Detection on Online Social Media using CNN with Pre-Trained Model

Abstract: Nowadays attackers move to image spam techniques instead of text based. The Majority of conventional techniques solely possess the capability to identify spam confined to textual content and hyperlinks. In this research we have done “Deep Convolutional Neural Networks” (DCNNs)in conjunction with pre-trained architectures. Image classification is one of the areas that has increased in the last decade very rapidly. But due to the less computational resources it become very challenging to train a good image classification model. With the help of Transfer learning, we can overcome this type of situation and can build a good image classification model. Our proposed model utilizes deep convolutional neural networks (DCNNs) along with pre-trained architectures. By fine-tuning this pre-trained model on the specific image dataset, the model can effectively learn to detect image- based spam content. Our suggested model surpasses contemporary state-of-the-art detection models concerning both accuracy and efficiency in the realm of the image spam identification.

Enhanced Detection of Text and Image Spam Using Cost-Sensitive Deep Learning

Enhanced Detection of Text and Image Spam Using Cost-Sensitive Deep Learning

Survey Paper on Implementation of Abusive Language Detection in Chatting Application

Communication has become stronger due to exponential increase in the usage of social media in the last few years. People use them for communicating with friends, finding new friends, updating any important activities of their life, etc. Due to their growing popularity and deep reach, these mediums are infiltrated with huge Volume of spam messages. Spam message randomly sent to multiple addressees by all sorts of groups, but mostly lazy advertisers and criminals who wish to lead you to phishing sites. In this we are using various machine learning techniques for detecting spam in the short text messages and also Google vision API for detecting spam images. Key Words: Spam Filtration, Google Vision API, OWASP, Naive baise classifier, Dictionary Based Algorithm

Survey Paper on Implementation of Abusive Language Detection in Chatting Application

Communication has become stronger due to exponential increase in the usage of social media in the last few years. People use them for communicating with friends, finding new friends, updating any important activities of their life, etc. Due to their growing popularity and deep reach, these mediums are infiltrated with huge volume of spam messages. Spam message randomly sent to multiple addressees by all sorts of groups, but mostly lazy advertisers and criminals who wish to lead you to phishing sites. In this we are using various machine learning techniques for detecting spam in the short text messages and also Google vision API for detecting spam images. Key Words: Spam Filtration, Google Vision API, OWASP, Naive baise classifier, Dictionary Based Algorithm

Application of error level analysis in image spam classification using deep learning model

Image spam is a type of spam that contains text information inserted in an image file. Traditional classification systems based on feature engineering require manual extraction of certain quantitative and qualitative image features for classification. However, these systems are often not robust to adversarial attacks. In contrast, classification pipelines that use convolutional neural network (CNN) models automatically extract features from images. This approach has been shown to achieve high accuracies even on challenge datasets that are designed to defeat the purpose of classification. We propose a method for improving the performance of CNN models for image spam classification. Our method uses the concept of error level analysis (ELA) as a pre-processing step. ELA is a technique for detecting image tampering by analyzing the error levels of the image pixels. We show that ELA can be used to improve the accuracy of CNN models for image spam classification, even on challenge datasets. Our results demonstrate that the application of ELA as a pre-processing technique in our proposed model can significantly improve the results of the classification tasks on image spam datasets.

Rare Category Analysis for Complex Data: A Review

Though the sheer volume of data that is collected is immense, it is the rare categories that are often the most important in many high-impact domains, ranging from financial fraud detection in online transaction networks to emerging trend detection in social networks, from spam image detection on social media platforms to rare disease diagnosis in medical decision support systems. The unique challenges of rare category analysis include (1) the highly skewed class distribution; (2) the non-separable nature of the rare categories from the majority classes; (3) data and task heterogeneity; and (4) the time-evolving property of the input data sources. This survey reviews state-of-the-art techniques used in complex rare category analysis, where the majority classes have a smooth distribution while the minority classes exhibit the compactness property in the feature space or subspace. Rare category analysis aims to identify, characterize, represent, and interpret anomalies that not only show statistical significance but also exhibit interesting patterns (e.g., compactness, high-order structures, showing in a burst). We introduce our study, define the problem setting, and describe the unique challenges of complex rare category analysis. We then present a comprehensive review of recent advances that are designed for this problem setting, from rare category exploration without any label information to rare category exposition that characterizes rare examples with a compact representation, from the representation of rare patterns in a salient embedding space to the interpretation the prediction results and providing relevant clues for the end-users’ interpretation. Finally we discuss potential challenges and shed light on the future directions for complex rare category analysis. 1

Attention Mechanism and Support Vector Machine for Image-Based E-Mail Spam Filtering

Spammers have created a new kind of electronic mail (e-mail) called image-based spam to bypass text-based spam filters. Unfortunately, these images contain harmful links that can infect the user’s computer system and take a long time to be deleted, which can hamper users’ productivity and security. In this paper, a hybrid deep neural network architecture is suggested to address this problem. It is based on the convolution neural network (CNN), which has been enhanced with the convolutional block attention module (CBAM). Initially, CNN enhanced with CBAM is used to extract the most crucial information from each image-based e-mail. Then, the generated feature vectors are fed to the support vector machine (SVM) model to classify them as either spam or ham. Four datasets—including Image Spam Hunter (ISH), Annadatha, Chavda Approach 1, and Chavda Approach 2—are used in the experiments. The obtained results demonstrated that in terms of accuracy, our model exceeds the existing state-of-the-art methods.

Deep learning-based spam image filtering

Spam is some unwanted material that may be put in the form of images. While many machine learning approaches are effective at detecting textual spam, this is not true for image spam. In this paper, a new framework for identifying image spams is proposed. The images are divided into two categories: undesirable material contained in the form of images which is referred to as a spam image, whereas anything else is referred to as a ham image. Our proposed methodology is based on applying different pre-trained deep learning models, including InceptionV3, Densely Connected Convolutional Networks 121(DenseNet121), Residual Networks (ResNet50), Visual Geometry Group (VGG16) and MobileNetV2, to filter out the unwanted spam images. Different standard test datasets such as Dredze Dataset, Image Spam Hunter (ISH) Dataset and Improved Dataset are utilized in this paper for performance testing. Furthermore, transfer learning and data augmentation are employed to address the issue of a shortage of labeled data. In our implementation, the fully connected (FC) layer in the aforementioned pre-trained models is replaced with a Support Vector Machine (SVM) classifier, resulting in an improved accuracy. The obtained results reveal that ResNet50 model yields the best performance achieving 99.87% accuracy, 99.88% area under the curve (AUC), 99.98% sensitivity, 99.79% precision, 98.99% F1 score and a computational testing time of in the order of one to two seconds for the ISH dataset.

Bi-modal email spam detection using recurrent and convolution neural network techniques

The increasing adoption of electronic emails as a means of communication, both at the commercial, government, and individual levels, serves as an impetus for attackers to compromise communication. Consequently, numerous machine learning techniques have been developed for identifying unwanted emails, commonly known as spam. Despite the significant progress reported in existing literature, most studies do not integrate the detection of both textual and image-based spam. In this paper, two deep learning techniques that detect both textual and image-based spam were evaluated. First, the Recurrent Neural Network (RNN) and Convolutional Neural Network (CNN) is studied, training them on various textimage features to explore their effectiveness on an improved dataset. Subsequently, in an effort to outsmart current spam detection techniques, a bi-modal architecture capable of detecting textual spam, image spam, and mixed spam is designed. The experimental results in conjunction with existing transfer learning for effective spam detection is provided.

An Improved Image Spam Classification Model Based on Deep Learning Techniques

Image Spam is a type of spam that has embedded text in an image. Classification of Image Spam is done using various machine learning approaches based on a broad set of features extracted from the image. For its remarkable results, the convolutional neural networks (CNN) are widely used in image classification as well as feature extraction tasks. In this research, we analyze image spam using a CNN model based on deep learning techniques. The proposed model is fine-tuned and optimized for both feature extraction as well as for classification tasks. We also compared our proposed model to different “Improved” and “Challenge” image spam datasets, which were developed for increasing the difficulty of the classification task. Our model significantly improves the accuracy of the classification task as compared to other approaches on the same datasets.

Hybrid Features by Combining Visual and Text Information to Improve Spam Filtering Performance

The development of information and communication technology has created many positive outcomes, including convenience for people; however, cases of unsolicited communication, such as spam, also occur frequently. Spam is the indiscriminate transmission of unwanted information by anonymous users, called spammers. Spam content is indiscriminately transmitted to users in various forms, such as SMS, e-mail, and social network service posts, causing negative experiences for users of the service, while also creating costs, such as unnecessarily large amounts of network traffic. In addition, spam content includes phishing, hype or false advertising, and illegal content. Recently, spammers have also used images that contain stimulating content to effectively attract users’ curiosity and attention. Image spam contains more complex information than text, making it more difficult to analyze and to generalize its properties compared to text. Therefore, existing text-based spam detectors are vulnerable to spam image attacks, resulting in a decline in service quality. In this paper, a “hybrid features by combining visual and text information to improve spam filtering performance” method is proposed to reduce the occurrence of misclassification. The proposed method employs three sub-models to extract features from spam images and a classifier model to output the results using the features. Each sub-model extracts topic-, word-, and image-embedding-based features from spam images. In addition, the sub-models use optical character recognition, latent Dirichlet allocation, and word2Vec techniques to extract features from images. To evaluate spam image classification performance, the spam classifiers were trained using the extracted features and the results were measured using a confusion matrix. Our model achieved an accuracy of 0.9814 and a macro-F1 score of 0.9813. In addition, the application of OCR evasion techniques resulted in a decrease in recognition performance. Using the proposed model, a mean macro-F1 score of 0.9607 was obtained.

Multivariate bounded support asymmetric generalized Gaussian mixture model with model selection using minimum message length

In this paper, the bounded support asymmetric generalized Gaussian mixture model (BAGGMM) is proposed for data modeling as an alternative to unbounded mixture models for the cases when the data lies in the bounded support region. The model parameters are learned through maximum likelihood estimation, and Expectation–Maximization (EM) with the Newton Raphson method is adopted for optimization of parameters. Model selection in mixtures is also considered an integral part of clustering; thus, we also have proposed a model selection criterion for BAGGMM through minimum message length. In order to validate the performance of the mixture model, it is applied to image spam detection, object clustering, and visual scene categorization. For the experiments, Spam Hunter, ETHZ, GHIM, and 15-Scene image datasets are adopted, and several clustering scenarios are developed to see the effectiveness of the proposed model. The clustering framework is also compared with AGGMM in a similar setting to all the experiments. In the next phase, the whole clustering framework is extended to examine the performance of the proposed model selection criterion and compared with different techniques to find the optimal mixture component, which further improves the clustering process. The experiments show that the proposed BAGGMM and model selection criterion has demonstrated its success in several learning applications.

Gallery Image Classification

The issue of peripheral device space management is a pressing concern for everyday users who deal with massive amounts of media content obtained through social media applications. The content that is shared over the social media application can be important images such as notes, personal pictures like selfies, group pictures, etc., and along with these images, there are many unsolicited images shared that are directly stored in the device. These unsolicited/spam images not only consume storage space but also selecting them from all and then deleting them is a tedious process, sometimes by mistake, a user also deletes important images, which can consume a lot of time. This project aims to specifically detect and extract study notes, personal images & spam images separately and relocate them in three different folders, which will help to segregate the images. This will make searching for a particular image easier and save time and effort. Keywords- Unsolicited images, notes images, personal images

Image SPAM Detection Using ML and DL Techniques

Abstract Since e-mail is one of the most common places to send messages, spammers have, in recent years, targeted it as a preferred way of distributing undesired messages (spam) to several users to spread viruses, cause destruction, and obtain user's information. Spam images are considered one of the known spam types. The spammer processes images and changes their characteristics, especially background colour, font type, or adding artefacts to the images to spread spam. In this paper, we proposed a spam detection model using Several ML (Random-Forest (RF), Decision-Tree (DT), KNearest Neighbor (KNN), Support-Vector Machine (SVM), NaïveBays (NB), and Convolutional Neural Network (CNN)). Several experiments evaluate the efficiency and performance of the (ML) algorithms for spam detection. Using the Image Spam Hunter Dataset extracted from real spam e-mails, the proposed model achieved over 99% accuracy on spam image detection. Keywords: SPAM, Machine Learning, Image Classification, Feature Extraction, Deep Learning.

Classification of Image Spam Using Convolution Neural Network

Image identification and classification is a basic issue in the fields of mainframe visualization and pattern recognition. In today’s world, a great deal of unwanted material is distributed via the Internet. The unwanted information contained inside images, i.e., image spam, endangers email-based communication systems. Unlike textural spam, image spam is difficult to be detected by many machine learning (ML) techniques. This paper intends to investigate and evaluate four deep learning (DL) methods that may be useful for image spam identification. Firstly, neural networks, especially deep neural networks, were trained on various image features. Their resilience was measured on an enhanced dataset, which was created specifically to outwit existing image spam detection methods. Next, a convolution neural network (CNN) was designed, and verified through experiments. Experimental results show that our novel approach for image spam identification outshines other current techniques in the field.

Development of a Machine Learning Model for Image-based Email Spam Detection

Combatting email spam has remained a very daunting task. Despite the over 99% accuracy in most non-image-based spam email detection, studies on image-based spam hardly attain such a high level of accuracy as new email spamming techniques that defeat existing spam filters emerges from time to time. The number of email spams sent out daily has remained a key factor in the continued use of spam. In this paper, a simple convolutional neural network model, 123DNet was developed and trained with 28,929 images drawn from 2 public datasets and a Personally Generated dataset. The model was optimized to the least set of layers to have 1 input layer, 2 embedded Convolutional layers as a hidden layer, and 3 neural network layers. The model was tested with a total of 4,339 images of the three dataset samples and then with a separate set of 1,200 images to test performance on never-seen-before images. A Classification Performance analysis was carried out using the confusion matrix. Performance metrics including Accuracy, Precision, True Negative Accuracy, Sensitivity, Specificity, and F1 Measure were computed to ascertain the model’s performance. The Model returned an F1 Score of 97% on a public dataset’s test sample and 88% on Never-seen-before test samples outperforming some pre-existing models while performing significantly well on the newly generated image test samples. It is recommended that a model that performed so well with new never-seen-before spam images be integrated into spam filtering systems. Keywords- Convolutional Neural Network, Deep Learning, Image-based Spam Detection

PROTECTOR: An optimized deep learning-based framework for image spam detection and prevention

Image spam is a spamming technique that integrates spam text content into graphical images in order to bypass conventional text-based spam filters. In order to detect image spam efficiently, it is important to analyze the image data. The existing image spam detection techniques in literature focus on textual or graphic features of the image. None of the existing techniques considered the link information of the image which results in low accuracy and performance degradation. So, to fill these gaps, in this paper, we analyzed the link properties of image, for image spam detection and prevention. We propose an optimized framework called as PROTECTOR. In PROTECTOR, the rank score is generated by using the linkage information of the image, textual information of the image, and metadata information of the image. The computed rank score indicates the relevance of an image. This rank score is then used to train a deep neural network design of deep learning, which yields the accuracy of 96% with respect to various performance evaluation metrics. Also, the optimization algorithm, i.e., genetic algorithm is fitted in the results according to the defined fitting function. The proposed framework is validated with standard dataset of Image spam Hunter.

A New Bio-Inspired Method for Spam Image-Based Emails Filtering

Electronic mail has become one of the most popular and frequently used channels for personal and professional online communication. Despite its benefits, e-mail faces a major security problem, which is the daily reception of a large number of unsolicited electronic messages, known as “spam emails.” Today, most electronic mail systems have simple spam filtering mechanisms based on text spam filtering technologies. To circumvent these filters, spammers are introducing new techniques of embedding spam messages in the image attached to the mail. In this article, the authors propose a new method for spam image filtering. The proposed system can distinguish between legitimate images from spam images based on the texture characteristics of the image attached to an email. From each image, around 20 characteristics can be extracted from the gray level co-occurrence matrix (GLCM). Then, to classify the images as spam or ham, the authors use a new metaheuristic nature-inspired model for building classifiers based on the social worker bees and enhanced nearest-centroid classification method.

Weighted k-Nearest Neighbour for Image Spam Classification

E-mail is an efficient and reliable data exchange service. Spams are undesired e-mail messages which are randomly sent in bulk usually for commercial aims. Obfuscated image spamming is one of the new tricks to bypass text-based and Optical Character Recognition (OCR)-based spam filters. Image spam detection based on image visual features has the advantage of efficiency in terms of reducing the computational cost and improving the performance. In this paper, an image spam detection schema is presented. Suitable image processing techniques were used to capture the image features that can differentiate spam images from non-spam ones. Weighted k-nearest neighbor, which is a simple, yet powerful, machine learning algorithm, was used as a classifier. The results confirm the effectiveness of the proposed schema as it is evaluated over two datasets. The first dataset is a real and benchmark dataset while the other is a real-like, modern, and more challenging dataset collected from social media and many public available image spam datasets. The obtained accuracy was 99.36% and 91% on benchmark and the proposed dataset, respectively.

FedLearnSP: Preserving Privacy and Security Using Federated Learning and Edge Computing

Enormous amount of information is processed at different websites, on a number of different AI tools and in multiple data silos. Sharing data between various sources, this is a significant obstacle, due to administrative, organizational, and security considerations. One possible solution is federated machine learning (FML), a system that simultaneously sends machine learning algorithms to all data sources, trains models at each source, and aggregates the learned models. This technique ensures consumer-influenced solution by processing the data locally. This work is the first to investigate the applicability of Internet attack detection through FML, to the best of our knowledge. Our primary contributions include the application of federated learning to satisfy customer search queries by detecting malicious spam images, which may lead these AI systems to retrieve irrelevant information. We assess and analyze the FML-entangled learning output comprehensively in different ways adjustments including balanced and imbalanced customer data distribution, scalability, and overhead communication. Our measuring results show that FML suits practical scenarios, where the variable image size, including the animation ratio to legitimate samples of images, present among the advertisements that may distract consumer from fetching relevant results. With the evaluated results, the state-of-the-art FedLearnSP proved significant image spam detection.