Weighted k-Nearest Neighbour for Image Spam Classification

Pattern Recognition Letters, 2011

In their arms race against developers of spam filters, spammers have recently introduced the image spam trick to make the analysis of emails' body text ineffective. It consists in embedding the spam message into an attached image, which is often randomly modified to evade signature-based detection, and obfuscated to prevent text recognition by OCR tools. Detecting image spam turns out to be an interesting instance of the problem of content-based filtering of multimedia data in adversarial environments, which is gaining increasing relevance in several applications and media. In this paper we give a comprehensive survey and categorisation of computer vision and pattern recognition techniques proposed so far against image spam, and make an experimental analysis and comparison of some of them on real, publicly available data sets.

Proceedings of the 8th International Workshop on Pattern Recognition in Information Systems, 2008

The presence of unsolicited bulk emails, commonly known as spam, can seriously compromise normal user activities, forcing them to navigate through mailboxes to find the-relatively few-interesting emails. Even if a quite huge variety of spam filters has been developed until now, this problem is far to be resolved since spammers continuously modify their malicious techniques in order to bypass filters. In particular, in the last years spammers have begun vehiculating unsolicited commercial messages by means of images attached to emails whose textual part appears perfectly legitimate. In this paper we present a method for overcoming some of the problems that still remain with state-of-the-art spam filters when checking images attached to emails. Results on both personal and publicly available email databases are presented, in order to assess the performance of the proposed approach.

2006

In recent years anti-spam filters have become necessary tools for Internet service providers to face up to the continuously growing spam phenomenon. Current server-side anti-spam filters are made up of several modules aimed at detecting different features of spam e-mails. In particular, text categorisation techniques have been investigated by researchers for the design of modules for the analysis of the semantic content of e-mails, due to their potentially higher generalisation capability with respect to manually derived classification rules used in current server-side filters. However, very recently spammers introduced a new trick consisting of embedding the spam message into attached images, which can make all current techniques based on the analysis of digital text in the subject and body fields of e-mails ineffective. In this paper we propose an approach to antispam filtering which exploits the text information embedded into images sent as attachments. Our approach is based on the application of state-of-the-art text categorisation techniques to the analysis of text extracted by OCR tools from images attached to e-mails. The effectiveness of the proposed approach is experimentally evaluated on two large corpora of spam e-mails.

Proceedings of the International Conference on Advances in Computer Science and Electronics Engineering, 2012

Today, the internet is the most powerful tools throughout the world. However the explosive growth of unsolicited emails has prompted the development of numerous spam filtering techniques It needlessly obstruct the entire system. Spammers are creating new ways against anti-spam technology. By the end of 2006, the nature of spam had totally shifted. The newest of which is the image-based spam. They are difficult to detect using traditional techniques. How to deal with the image spam is a difficult problem for the entire world. Image spam has become the main form of spam, it is a problem crying out for solutions to effectively filter such spam nowadays. This paper considers various systems that has been proposed for the detection of image spam by considering text, content and image features.

2008

In this paper we consider the approach to image spam filtering based on using image classifiers aimed at discriminating between ham and spam images, previously proposed by other authors. In previous works this approach was implemented using "generic" image features. In this paper we show that its effectiveness can be improved by using specific features related to the graphical characteristics of embedded text. The features we consider are derived from measures which were proposed in our previous works with the aim of detecting image obfuscation techniques often used by spammers to make OCR tools ineffective. An experimental investigation is carried out on a set of images taken from two corpora of real ham and spam emails.

2012

Today, the internet is the most powerful tools throughout the world. But the explosive growth of unsolicited emails has prompted the development of numerous spam filtering techniques It needlessly obstruct the entire system. Spammers are creating new ways against anti-spam technology. By the end of 2006, the nature of spam had totally shifted. The newest of which is the image-based spam. Traditional techniques are not sufficient to deal with these image spam. Combating with the image spam is a serious issue for the entire world. Image spam has become the main form of spam, it is a problem crying out for solutions to effectively filter such spam nowadays. In this paper a maximum likelihood method has been proposed for the classification of spam images.

2020

Today, images are a part of communication between people. However, images are being used to share information by hiding and embedding messages within it, and images that are received through social media or emails can contain harmful content that users are not able to see and therefore not aware of. This paper presents a model for detecting spam on images. The model is a combination of optical character recognition, natural language processing, and the machine learning algorithm. Optical character recognition extracts the text from images, and natural language processing uses linguistics capabilities to detect and classify the language, to distinguish between normal text and slang language. The features for selected images are then extracted using the bag-of-words model, and the machine learning algorithm is run to detect any kind of spam that may be on it. Finally, the model can predict whether or not the image contains any harmful content. The results show that the proposed method using a combination of the machine learning algorithm, optical character recognition, and natural language processing provides high detection accuracy compared to using machine learning alone.

AIRCC Publishing Corporation, 2019

Today very important means of communication is the e-mail that allows people all over the world to communicate, share data, and perform business. Yet there is nothing worse than an inbox full of spam; i.e., information crafted to be delivered to a large number of recipients against their wishes. In this paper, we present a numerous anti-spam methods and solutions that have been proposed and deployed, but they are not effective because most mail servers rely on blacklists and rules engine leaving a big part on the user to identify the spam, while others rely on filters that might carry high false positive rate.

To eliminate spam e-mails, several text anti-spam systems were created to analyze the textual content of e-mails and classify them. Owing to the good performance of these systems, spam messages began to occur in images. This rendered text anti-spam systems useless, thereby fostering the development of image anti-spam systems. Image processing is much more computationally expensive than text processing, and the results of image anti-spam systems have been inferior to those of text systems. This paper proposes an image anti-spam system that makes use of various methods of image feature extraction and an artificial neural model to classify e-mails. The extraction methods are evaluated both individually and in combination. The neural model is exhaustively evaluated using publicly available databases. The use of these databases is described in detail in order to facilitate reproducible results. Besides analyzing the classification ability of the proposed system, this study also evaluates its computational costs, including costs for extracting features and classifying images. The results are promising both in terms of rates of correct classification and of false positives produced by the anti-spam system, as well as in terms of its computational cost.

2007

Abstract Recently, spammers have proliferated “image spam”, emails which contain the text of the spam message in a human readable image instead of the message body, making detection by conventional content filters difficult. New techniques are needed to filter these messages. Our goal is to automatically classify an image directly as being spam or ham. We present features that focus on simple properties of the image, making classification as fast as possible.