A new methodology for detection of counterfeit Viagra® and Cialis® tablets by image processing and statistical analysis

https://doi.org/10.1016/j.forsciint.2011.09.002Get rights and content

Abstract

This paper proposes a new approach for automatic classification of counterfeit Viagra® and Cialis® tablets using image processing and statistical analysis. A high resolution VSC 5000 is used for image acquisition in a controlled environment, and the combination of a thresholding technique with morphological operators is used to segment the tablet from the background. A statistical model based on the RGB color components of original samples is built, and the detection of counterfeit tablets was performed by checking the adherence of a test sample to the obtained distribution using the Bhattacharyya distance. Our experimental results indicated that counterfeit tablets can be effective detected using the proposed approach.

Introduction

The production of counterfeit medicines is a criminal problem that carries serious risks to public health worldwide. Studies suggest that significant growth in the last decade may be associated with easier access by the counterfeiters to the technologies needed to copy genuine pharmaceutical products [1]. Moreover, marketing by the internet allows anyone to buy any medicine without prescription, easily and anonymously, sometimes from fraudulent sites [2], [3]. In an attempt to combat this phenomenon, the World Health Organization (WHO) created in 2006 a global coalition of stakeholders called IMPACT (International Medical Products Anti-Counterfeiting Taskforce). The taskforce has been active in forging international collaboration to seek global solutions to this global challenge and in raising awareness of the dangers of counterfeit medical products [4].

By definition, counterfeit medicines are pharmaceutical products that have “been deliberately and fraudulently mislabeled with respect to identity and/or source” [5]. All kinds of medicines are counterfeited [6], [7], [8], [9], [10], [11], [12], but the market success of the three approved phosphodiesterase type 5 (PDE-5) inhibitors for treating erectile dysfunction, sildenafil (Viagra®, Pfizer), tadalafil (Cialis®, Eli Lilly), and vardenafil (Levitra®, Bayer) has led to an explosion in counterfeit versions of these products [13]. PDE-5 inhibitors are a prime target for counterfeiting because of their high cost and the embarrassment associated with the underlying condition leading people to turn to the internet to buy these medicines easily, anonymously and often cheaply [13].

Following this trend, in Brazil, Viagra® and Cialis® are among the most counterfeited medicines. Routinely, seizures of suspected counterfeit medicines are forwarded to the Brazilian Federal Police (PF) for forensic analysis. The database of the PF show that from January 2007 to September 2010, 371 reports involving counterfeit medicines were issued, and of these, 295 (80%) included counterfeit Cialis® and/or Viagra®. As with other illegal products seized by the police, these figures should be just a small sample of the real market, i.e., it is the “tip of the iceberg”.

The verification of a counterfeit medicine is based primarily on a comparison between authentic and questioned products, involving detailed analysis of different elements in the existing packaging, drug leaflet and both the exterior and interior of the pharmaceutical dosage form. Regarding the analysis of the package, one could check for the presence of reactive ink on the packaging (that shows a text to be rubbed with metallic object), the existence of holographic security labels on packages, or recognize the print patterns on the package, including size and types of sources and figures. Such analysis is quite subjective, and clearly useless when the package is not available.

Regarding the analysis of the tablet itself, different advanced techniques have been proposed for the detection of counterfeits of Cialis® and/or Viagra® in recent years. Most are based on destructive chemical tests [2], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], while a few others are based on non destructives tests that make use of imaging technologies outside the visible range [13], [25], [126], requiring specific (and usually expensive) equipments. Recently, Rodomonte et al. [3] proposed an approach for counterfeit medicine detection by measuring the reflectance of tablets and package along the visible spectrum using a spectrophotometer. However, none of the methods is based on simple photographs of Cialis® and Viagra® tablets, which are easy to obtain and reflect global properties of the coat film.

The outer layer that covers the core of each tablet is responsible for color, uniformity and brightness characteristic for each coat. It is known that the pharmaceutical industry applies high-end technologies in the coating process plus a rigid quality control that discards any tablet with physical imperfections [27], so that a clandestine production is expected to generate coating patterns distinguishable from the original coating of tablets. Hence, our claim is that counterfeit tablet detection can be performed by analyzing color and homogeneity characteristics of the tablet, which can be obtained using visible light and a common optical camera.

Although any optical camera could be used to acquire images under visible light, the exact same tablet may look considerably different depending on the cameras lenses, sensor, and particularly lighting conditions. The Video Spectral Comparator – VSC – is a device used in the scientific examination of documents [28], [29], [30], [31], and it is available in almost all PF units. In particular, it allows the acquisition of images under controlled illumination conditions and using the exact same lens/sensor pair, being a good choice to evaluate counterfeit tablets.

Considering the points raised above, this paper presents a new analytical methodology for quickly, without sample preparation, non-destructively and reliably discrimination of genuine and fake Viagra® and Cialis® tablets, using images obtained with the VSC. The proposed approach consists of initially segmenting the tablet from the background image, and obtaining a statistical model of the RGB color components of genuine samples. The statistical distribution of a tablet being tested is then compared with the distribution of genuine tablets, and a counterfeit is detected if the dissimilarity exceeds a given threshold. As far as we know, this is the first approach for detecting counterfeit coated tablets based only on image processing techniques.

Section snippets

Image acquisition

Fake and authentic tablets [32] were used to generate a database of Viagra® and Cialis® samples. Authentic tablets were provided by Pfizer Ltda and Eli Lilly do Brasil Ltda, while counterfeit tablets were obtained from seizures forwarded to the Brazilian Federal police. A high resolution VSC 5000 (Foster & Freeman Ltd., UK) was used for image acquisition, and all images were captured using light in the visible region, with a magnification of 20× and automatic adjustment of brightness and

Experimental results

We have tested the proposed approach using two datasets: one containing 19 genuine and 24 counterfeit Viagra® tablets, and the other containing 20 genuine and 53 counterfeit Cialis® tablets. As explained in Section 2.3, validation was performed, for each dataset, using the leave-p-out procedure for different values of p.

Results were evaluated quantitatively in terms of the True Positive Rate TPR (or sensitivity) and the True Negative Rate TNR (or specificity), given byTPR=TPP,TNR=TNN,where TP

Conclusions

This paper proposed a new approach for automatic classification of counterfeit coated tablets using image processing and statistical analysis, focusing on Viagra® and Cialis® tablets. A high resolution VSC 5000 was used to acquire an image of each tablet in a controlled environment. A simple binarization approach followed by morphological post-processing was used to segment the tablet from the background, and the statistical distribution of color pixels related to genuine tablets was computed.

Acknowledgement

Author Claudio Jung would like to thank Brazilian funding agency CNPq for supporting this work.

References (36)

Cited by (37)

  • Identifying the most relevant tablet regions in the image detection of counterfeit medicines

    2021, Journal of Pharmaceutical and Biomedical Analysis
    Citation Excerpt :

    One way to deal with this threat is to develop and improve analytical techniques to effectively detect counterfeiting [3,4]. The set of analytical techniques traditionally used to analyze counterfeit drugs is wide, ranging from the assessment of tablets physical profile, X-ray fluorescence, Near Infrared (NIR) spectroscopy and hyperspectral imaging to visible light imaging [5,6]. Among such analytical techniques, visible light imaging stands out as a non-destructive and reliable analytical technique that dismisses sample preparation and provides fast results [6].

  • Applications of machine vision in pharmaceutical technology: A review

    2021, European Journal of Pharmaceutical Sciences
View all citing articles on Scopus
View full text