Imperceptible reversible watermarking of radiographic images based on quantum noise masking.

BACKGROUND AND OBJECTIVE: Advances in information and communication technologies boost the sharing and remote access to medical images. Along with this evolution, needs in terms of data security are also increased. Watermarking can contribute to better protect images by dissimulating into their pixels some security attributes (e.g., digital signature, user identifier). But, to take full advantage of this technology in healthcare, one key problem to address is to ensure that the image distortion induced by the watermarking process does not endanger the image diagnosis value. To overcome this issue, reversible watermarking is one solution. It allows watermark removal with the exact recovery of the image. Unfortunately, reversibility does not mean that imperceptibility constraints are relaxed. Indeed, once the watermark removed, the image is unprotected. It is thus important to ensure the invisibility of reversible watermark in order to ensure a permanent image protection. METHODS: We propose a new fragile reversible watermarking scheme for digital radiographic images, the main originality of which stands in masking a reversible watermark into the image quantum noise (the dominant noise in radiographic images). More clearly, in order to ensure the watermark imperceptibility, our scheme differentiates the image black background, where message embedding is conducted into pixel gray values with the well-known histogram shifting (HS) modulation, from the anatomical object, where HS is applied to wavelet detail coefficients, masking the watermark with the image quantum noise. In order to maintain the watermark embedder and reader synchronized in terms of image partitioning and insertion domain, our scheme makes use of different classification processes that are invariant to message embedding. RESULTS: We provide the theoretical performance limits of our scheme into the image quantum noise in terms of image distortion and message size (i.e. capacity). Experiments conducted on more than 800 12 bits radiographic images of different anatomical structures show that our scheme induces a very low image distortion (PSNR∼ 76.5 dB) for a relatively important capacity (capacity∼ 0.02 bits of message per pixel). CONCLUSIONS: The proposed watermarking scheme, while being reversible, preserves the diagnosis value of radiographic images by masking the watermark into the quantum noise. As theoretically and experimentally established our scheme offers a good capacity/image quality compromise that can support different watermarking based security services such as integrity and authenticity control. The watermark can be kept into the image during the interpretation of the image, offering thus a continuous protection. Such a masking strategy can be seen as the first psychovisual model for radiographic images. The reversibility allows the watermark update when necessary.

Multiple-Instance Learning for Anomaly Detection in Digital Mammography.

This paper describes a computer-aided detection and diagnosis system for breast cancer, the most common form of cancer among women, using mammography. The system relies on the Multiple-Instance Learning (MIL) paradigm, which has proven useful for medical decision support in previous works from our team. In the proposed framework, breasts are first partitioned adaptively into regions. Then, features derived from the detection of lesions (masses and microcalcifications) as well as textural features, are extracted from each region and combined in order to classify mammography examinations as "normal" or "abnormal". Whenever an abnormal examination record is detected, the regions that induced that automated diagnosis can be highlighted. Two strategies are evaluated to define this anomaly detector. In a first scenario, manual segmentations of lesions are used to train an SVM that assigns an anomaly index to each region; local anomaly indices are then combined into a global anomaly index. In a second scenario, the local and global anomaly detectors are trained simultaneously, without manual segmentations, using various MIL algorithms (DD, APR, mi-SVM, MI-SVM and MILBoost). Experiments on the DDSM dataset show that the second approach, which is only weakly-supervised, surprisingly outperforms the first approach, even though it is strongly-supervised. This suggests that anomaly detectors can be advantageously trained on large medical image archives, without the need for manual segmentation.

Multiple-instance learning for breast cancer detection in mammograms.

This paper describes an experimental computer-aided detection and diagnosis system for breast cancer, the most common form of cancer among women, using mammography. The system relies on the Multiple-Instance Learning (MIL) paradigm, which has proven useful for medical decision support in previous works from our team. In the proposed framework, the breasts are first partitioned adaptively into regions. Then, either textural features, or features derived from the detection of masses and microcalcifications, are extracted from each region. Finally, feature vectors extracted from each region are combined using an MIL algorithm (Citation k-NN or mi-Graph), in order to recognize "normal" mammography examinations or to categorize examinations as "normal", "benign" or "cancer". An accuracy of 91.1% (respectively 62.1%) was achieved for normality recognition (respectively three-class categorization) in a subset of 720 mammograms from the DDSM dataset. The paper also discusses future improvements, that will make the most of the MIL paradigm, in order to improve "benign" versus "cancer" discrimination in particular.

A joint encryption/watermarking system for verifying the reliability of medical images.

In this paper, we propose a joint encryption/water-marking system for the purpose of protecting medical images. This system is based on an approach which combines a substitutive watermarking algorithm, the quantization index modulation, with an encryption algorithm: a stream cipher algorithm (e.g., the RC4) or a block cipher algorithm (e.g., the AES in cipher block chaining (CBC) mode of operation). Our objective is to give access to the outcomes of the image integrity and of its origin even though the image is stored encrypted. If watermarking and encryption are conducted jointly at the protection stage, watermark extraction and decryption can be applied independently. The security analysis of our scheme and experimental results achieved on 8-bit depth ultrasound images as well as on 16-bit encoded positron emission tomography images demonstrate the capability of our system to securely make available security attributes in both spatial and encrypted domains while minimizing image distortion. Furthermore, by making use of the AES block cipher in CBC mode, the proposed system is compliant with or transparent to the DICOM standard.