Search

Non-intrusive and automated detection of pig breeds, particularly from visual standpoint, is important from a food quality tracking perspective. In this work, colour as well as texture based visual descriptors from muzzle images have been identified, which, serve as breed-identifiers to separate four common pig-breeds: Duroc, Ghungroo, Hampshire and Yorkshire. While these handcrafted visual descriptors by themselves are fairly robust and discriminative, it is recognized that by controlling the decision space by choosing the feature-type based on colour or texture or both and the order in which particular breeds are siphoned, classification accuracy can be improved considerably. In that light, a stable, relatively data-independent, breed-specific, hierarchical tree synthesis and feature selection procedure is proposed based on a breed-pair cluster separation table. The proposed approach has been compared with the state of the art Phylogenetic distance based Hierarchical Agglomerative Clustering algorithm (AGNES) and also with the standard decision tree classification algorithm. On cross-validation, When completely different sets of pigs were used for training and testing (50-50 split), the proposed algorithm reported relatively high mean classification accuracies of 86.45% for Duroc, 93.02% for Ghungroo, 86.91% for Hampshire and 98.54% for Yorkshire, respectively.

Existing architectures used in face anti-spoofing tend to deploy registered spatial measurements to generate feature vectors for spoof detection. This means that the ordering or sequence in which specific statistics are computed cannot be changed, as one moves from one facial profile to another. While this arrangement works in a person-specific setting, it becomes a major drawback when single-sided training is done based on the natural face class alone. To mitigate subject identity linked content interference within the anti-spoofing frame, we propose a identity-independent architecture based on random correlated scans of natural face images. The same natural face image can be scanned multiple times through independent correlated random walks before deriving simple differential features on the 1D scanned vectors. This proposed frame tends to capture the pixel correlation statistics with minimal content interference and shows great promise, particularly when trained on natural face sets, using a one-class support vector machine and cross-validated on other databases. Performance measured in terms of EER for detection of spoof face is found to be $$3.8291\%$$ with proposed random scan features, and $$2.02\%$$ with auto-population samples for inter database.

In natural photography, defects in camera imaging pipeline often result in some form of colour noise or distortion. Nature of this distortion is generally intertwined with scene dependent variables such as positioning, intensity and composition of light source and local object colour reflectivity. One such defect is called purple fringing aberration (PFA). PFA problems are of two types, one of which corresponds to a localised fringing effect near high contrast zones (termed as Isolated PFA or IS-PFA) and the second which corresponds to a widespread semi-transparent purple haze over a large part of natural scene (termed as complex PFA or C-PFA). Much of the PFA-correction solutions have been driven towards IS-PFA and very little towards C-PFA. Based on a premise that in C-PFA, green channel is heavily suppressed and noisy, while colour information in red and blue channels are largely conserved, authors propose a green-channel compensation algorithm for restoring true natural colours in fringe affected region. To correct white-tuft produced by proposed compensation algorithm, they also devise a suitable localised luminance adaptation procedure to equalise perceived changes in luminance profile. Comparisons with state-of-the-art methods devised to combat this purple haze effect yield promising results for a majority of test cases.

Quality assessment of Purple Fringing Aberrated (PFA) images remains an unsolved problem because the original untainted natural image of the scene is not available at the point of analysis. As a result, the problem assumes the form of a blind assessment. This PFA is a false coloration localized around the edge regions where the contrast differential is high. One can, therefore, surmise that if this coloration is largely homogeneous in the chrominance space, the edge is expected to be crisp and the image sharp and clear. However, if this coloration pattern is diverse in the chrominance space, the edges will be fuzzy and this, in turn, will have an impact on the visual clarity of the image. The fringe diversity, therefore, becomes a measure of PFA image quality, provided the fringes are distributed in several parts of the image. This diversity has been captured by first characterizing the chrominance space spanned by the PFA pixels and then using the eigenvalue ratio as a measure of color diversity.

The need for facial anti-spoofing has emerged to counter the usage of facial prosthetics and other forms of spoofing at unmanned surveillance stations. While some part of literature recognizes the difference in texture associated with a prosthetic in comparison with a genuine face, the solutions presented are largely prosthetic model specific and rely on two-sided calibration and training. In this paper, we focus on the specular component associated with genuine faces and claim that on account of the natural depth variation, its feature diversity is expected to be much larger as compared to prosthetics or even printed photo impersonations. In our work concerning one-sided calibration, we first characterize the specular feature space corresponding to genuine images and learn the projections of genuine and spoof data onto this basis. The trained SVM corresponding to genuine projections, 3D mask projections, and printed photo projections is then used as an anti-spoofing model for detecting impersonations.

Visual features associated with cattle and pigs cannot just be used for individual biometric identification, they can also be used to organize the domain into distinct breeds. In this paper, we use cropped muzzle images of pigs for breed identification. Gradient patch density maps are first created, and then the patch density profile distribution tailored to a specific breed is learnt to characterize the feature space for each of the four pig breeds: Duroc, Ghungroo, Hampshire, and Yorkshire. A Maximal Likelihood (ML) inferencing at the patch level followed by a second-tier decision fusion based on majority vote has been used to detect the breed from any query feature computed from an arbitrary muzzle image. Duroc, Ghungroo, and Yorkshire show good classification accuracies of 75.86%, 70.59%, and 100%, respectively, while the accuracy drops for Hampshire to 58.78% on account of the intrinsic white patch diversity in the breed and its similarity to Duroc and Ghungroo on some counts.

Conventional face anti-spoofing paradigms tend to operate on plain facial profiles and learn either the natural face space alone (one-class training problem) or both the natural face space as well as the spoof sample space (2-class training problem). However, this rigidity with respect to spatially constrained measurements, makes the base feature or statistic vulnerable to noise related to pose and camera perspective/orientational and scale changes. Noting that the sharpness profile computed on a natural face is largely independent of the pose and perspective change, it is imperative that the measurements be extracted in an identity independent setting by ignoring the pose/perspective variation. To facilitate this, we have deployed a 2-dimensional random walk for capturing lower order pixel correlation statistics from natural faces, with virtually no perceptual interference. The proposed identity independent frame has surpassed the state of the art with reference to a 3D mask dataset (image oriented, isolated frame setting), with an EER of \(2.25\%\) without auto-population and an EER of \(0.45\%\) with auto-population.

Precursor B-cell lymphoblastic lymphoma presenting with severe pain and extensive skeletal lesions as an initial presentation is very rare. We report a nine year old boy who presented with lower backache and difficulty in walking of 2 months duration. He neither had lymphadenopathy nor organomegaly. His peripheral blood picture and bone marrow examination were normal. MRI revealed multiple lytic lesions in the spine and PET scan revealed metabolically active lytic lesions in the axial skeletal system and bilateral femurs and humerus. Biopsy of the vertebra revealed the diagnosis of precursor B cell lymphoma. A diagnosis of malignancy should never be missed in a patient with musculoskeletal pain, having ruled out infectious and traumatic etiologies especially when they present with several skeletal radiolucencies.

Purple fringe aberration (PFA) patterns stem from specific defects in certain camera sensor grids, leading to the fraying of edges near high-contrast regions. Much of the literature deploys predefined, absolute, experimentally determined static thresholds for detecting purple fringes. Given the potential diversity in the spectral signature of the local light source, these fringes may not have a static distribution. It is therefore important to make the detection procedure content-adaptable and in tune with the environmental settings. In this paper, we propose a PFA detection procedure in the Y-Cb-Cr CHROMATIC space, by first using a global relativistic Y-channel-gradient threshold for detecting the high-contrast regions and then use the fact that PURPLE and GREEN are antipodes of each other to segregate PURPLE fringes reliably. Comparisons with the state-of-the-art detection approaches are presented. The advantage with the proposed approach rests with the fact that the threshold is content-adaptable and non-static and can therefore be used to pick up diverse fringe patterns (not necessarily confined to the seat of PURPLE).

Natural images particularly connected with portraits exhibit a diversity in the sharpness profile commensurate with varying depth associated with different parts of the face. This depth variation puts the object (the 3D face) slightly out of focus in different parts of the image. When a secondary image is created from this portrait, this blurring effect is amplified particularly if the plane of the printed photo is not aligned with the object plane of focus. Thus photos of photos exhibit greater homogeneity in the sharpness profiles and have overall lower sharpness values than their natural counterparts. This principle has been demonstrated through a lens model and has been applied towards face anti-spoofing. Proposed system was tested on the CASIA dataset and showed a recognition rate of 98.38% corresponding to a false positive rate of 10%.

In a cloud environment to preserve privacy of the end user, there arises a need to match and classify images in the encrypted domain. To impart granularity to image-data access it is important to make the matching process key independent, without compromising on the strength of the cipher or precision in the matching process. In this paper we propose a simple block intensity substitution as an encryption algorithm. The feature that remains transparent to this form of encryption is the sorted intensity histogram and its derivatives. These secondary and tertiary histogram based features remain robust to pose variations and yet discriminating across different subjects. A blind encrypted image classification procedure based on a standard unsupervised clustering algorithm has been used to validate the proposed concept.

This paper presents a completely blind or no-reference metric for estimation of perceived noise in images. This novel metric dubbed CINEMA (Content Independent Noise Estimation for Multimedia Applications) is completely content unaware and aligns well with human perception. An HOG-based model is employed for selection of weak textured patches and a wavelet decomposition strategy is used for detecting and quantifying noise. Experimental results on the LIVE database show that CINEMA achieves consistently good performance for different noise levels as compared to many of the existing Full Reference and No Reference Image Quality Assessment (IQA) metrics. We further show how CINEMA can be used to obtain an estimate (σ) of the noise standard deviation (σ) with high accuracy. A MATLAB implementation of the model is available at https://sites.google.com/site/blindiqa/cinema.

Planar spoofing is a well researched problem, wherein a high quality planar photograph can be replayed in front of a still camera as a substitute for another individual's face. Most modern day face recognition systems can be fooled by this process, as the perceptual information contained in a photo-of-a-photo, is virtually the same as that of a natural photograph of an individual. Current solutions attempt to detect this form of planar-spoofing through an extrinsic training process wherein both planar samples as well as regular photos are included as separate training sets. To avoid this form of explicit discriminant model-learning, we propose a single class training procedure for establishing and quantifying the quality of natural photographs taken under different lighting conditions, in terms of their CONTRAST PROFILE. Once this distribution is learnt, a suitable threshold is set based on the mean and standard deviation to pick up outliers. In this paper, we show that with just single poses of subjects, it is possible to achieve a low Equal Error Rate (EER) of 21.56% on the CASIA dataset and a rate of 8.57% upon cross-validation with a trimmed and shortened version of the MSU dataset.

Purple Fringes are aberrations or defects in the camera sensor profile which are spotted in the vicinity of edges near high contrast regions. Since these defects are very device centric, they become potential Forensic Tools for detection and identification of splicing operations in natural images. In existing literature, these fringe patterns are localized and mapped to what are known as Directive PFA-events. It was claimed in existing literature, that this swarm of directional PFA-events will exhibit a convergence towards the image centre for authentic images and a divergence for spliced images. By re-creating the PFA-event synthesis process for both authentic and forged images, we demonstrate both with rational arguments and experimental evidence that this convergence is not possible. This dismisses the PFA-event based approach as a Forensic Tool, without ruling out other secondary features which may be derived from the fringe regions for forensic analysis.

Search through a database of encrypted images against a crumpled and encrypted query will remain privacy preserving only if comparisons between selective features derived from these images is executed in the encrypted domain itself. To facilitate this, the encryption process must remain transparent to specific image statistics computed in the spatial or transform domain. Consequently, the perceptual hash formed by quantizing the image statistics remains the same before and after the encryption process. In this paper, we propose a transparent privacy preserving hashing scheme tailored to preserve the DCT-AC coefficient distributions, despite a constrained inter-block shuffling operation. These DCT distributions can be mapped onto a generalized Gaussian model characterized by shape and scale parameters, which can be quantized and Gray-coded into a binary hash matrix. The encryption scheme has been shown to be perceptually secure and does not impair the search reliability and accuracy of the hashing procedure. Experimental results have been provided to verify the robustness of the hash to content-preserving transformations, while demonstrating adequate sensitivity to discriminate between different images.

ion of a fingerprint in the form of a hash can be used for secure authentication. The main challenge is in finding the right choice of features which remain relatively invariant to distortions such as rotation, translation and minutiae insertions and deletions, while at the same time capturing the diversity across users. In this paper, an alignment-free novel fingerprint hashing algorithm is proposed which uses a graph comprising of the inter-minutia minimum distance vectors originating from the core point as a feature set called the minimum distance graph. Matching of hashes has been implemented using a corresponding search algorithm. Based on the experiments conducted on the FVC2002-DB1a and FVC2002-DB2a databases, we obtained an equal error rate of 2.27%. The computational cost associated with our fingerprint hash generation and matching processes is relatively low, despite its success in capturing the minutia positional variations across users.

Static storage of decryption keys in RFID tags creates a security issue particularly when some of these tags are compromised. To address this problem, a framework is proposed in which these tags search and compute decryption keys based on specific intrinsic association rules embedded in the tags, which feed on publicly known broadcast messages transmitted by the centre. This association rule is nothing but a circular linked list which connects a set of T tokens, in some random order. To facilitate backward secrecy, we also propose a rule adaptation methodology based on random deletions within this circular linked list triggered by random numbers, sent by the centre. We have shown theoretically that the search space for tags in possession of the actual keys is linear in the number of tokens contained in the association rule i.e. O(T ), while the search space for eavesdropping tags increases considerably to O(Tr), where r is centre-defined as the length of the footprint, within a circular linked list. Tradeoffs which involve balancing the extent of backward secrecy with network lifetime, are discussed.

Morphing, fusion and stitching of digital photographs from multiple sources is a common problem in the recent era. While images may depict visual normalcy despite a splicing operation, there are domains in which a consistency or an anomaly check can be performed to detect a covert digital stitching process. Most digital and low-end mobile cameras have certain intrinsic sensor aberrations such as purple fringing (PF), seen in image regions where there are contrast variations and shadowing. This paper proposes an approach based on Fuzzy clustering to first identify regions which contain Purple Fringing and is then used as a forensic tool to detect splicing operations. The accuracy of the Fuzzy clustering approach is comparable with the state-of-the-art PF detection methods and has been shown to penetrate standard interpolation and stitching operations performed using ADOBE PHOTOSHOP.

Extracting film songs from a multilingual database based on a query clip is a challenging task. The challenge stems from the subtle variations in pitch and rhythm, which accompany the change in the singer’s voice, style, and orchestration, change in language and even a change in gender. The fingerprinting algorithm must be designed to capture the base tune in the composition and not the adaptations (or variations which include lyrical modifications and changes in the singer’s voice). The SHAZAM system was developed for capturing cover audio pieces from millions of Western songs stored in the database, with the objective of tapping into the melodic construct of the song (devoid of other forms of embellishments). When applied to the Indian database the system was found less effective, due to subtle changes in both rhythm and melody mainly due to the semiclassical nature of Indian film songs. The retrieval accuracy was found to be 85 %. Potential reasons for the failure of this SHAZAM system have been discussed with examples.

This paper provides a tutorial and survey of digital fingerprinting and video scrambling algorithms based on partial encryption. Necessary design tradeoffs for algorithm development are highlighted for multicast communication environments. We also propose a novel architecture for joint fingerprinting and decryption that holds promise for a better compromise between practicality and security for emerging digital rights management applications.

Conventionally eigen-decompositions based on Principal Component Analysis and its variations have been used as a learning tool for capturing the pose and illumination changes in a large set of images, particularly faces. However, if this eigen-decomposition is performed on a single face image based on the row-column covariance statistics, the resulting dominant eigenvectors can be used for checking the statistical-synchronicity between any two images. This comparison can be done by determining the degree of alignment between the dominant eigenvectors which span the row or column spaces in the two images. This eigen-linking process has been found to be robust to several signal processing operations, scaling and noise insertion, despite remaining sufficiently discriminative across perceptually dissimilar images.

Threshold ramp secret sharing schemes are designed so that (i) certain subsets of shares have no information about the secret, (ii) some subsets have partial information about the secret and (iii) some subsets have complete information to recover the secret. However most of the ramp schemes in present literature do not control the leakage of information in partial access sets, due to which the information acquired by these sets is devoid of structure and not useful for fine-grained access control. Through a non-perfect secret sharing scheme called MIX-SPLIT, an encoding methodology for controlling the leakage in partial access sets is proposed and this is used for fine-grained access to binary strings. The ramp code generated using MIX-SPLIT requires a much smaller share size of O(n), as compared to Shamir's ramp adaptation which incurs a share size of atleast O(n 2 ) for the same multi-access structure. The proposed ramp code is finally applied towards the protection and fine-grained access of industrial design drawings.

Secure multicast in a wireless network is possible only if subsets of nodes share different sets of encryption keys. In this paper we propose a key protection and release mechanism based on broadcast tokens, where each node is equipped with protected shares of several encryption keys (node-share) which can be unlocked only when the broadcast tokens arrive. If the node-shares are different for different nodes, each broadcast token could unlock a different set of keys in each node. This makes the network reconfigurable as the information shared between various nodes will change with the accumulation of each new token. A non-perfect secret sharing scheme has been used construct the node-shares and tokens based on a carefully designed codebook which must satisfy certain rules to ensure that the un-encrypted broadcast tokens and the node-shares do not leak the encryption keys. Construction for a three node re-configurable network is discussed under collusion-free conditions.

Background:Burst fractures are common injuries of dorsolumbar spine. In indicated cases, surgery is the treatment of choice. Significant controversy exists regarding surgical intervention for these fractures. Posterior decompression, anterior decompression and instrumentation, and combined anterior decompression and posterior instrumentation have been recommended in various studies. Here we are going to evaluate unstable burst fractures of thoracic and lumbar spine treated by isolated anterior decompression and instrumented fusion with TSM-Bone graft composite.Methods: Prospective study of thirty-six cases of unstable fracture of thoracic and lumbar spine treated in Sri Ramachandra Medical centre from January 2011 to January 2014. The inclusion criteria were burst fractures of thoracic or lumbar spine complete or incomplete neurological deficit and burst fractures of thoracic or lumbar spine without neurological deficit but with mechanical instability. The exclusion criteria were pathological fractures, chance fracture, stable burst, wedge compression and osteoporotic compression fractures. The results were analyzed during the follow-up using the Pain – Visual analogue scale, Fusion status and radiographic parameter – K-angle .For pain score were given as 3,2,1 for absent, moderate and severe pain respectively. Regarding fusion status score of 3,2,1 were given when fusion was good, fair and no sign of fusion respectively.Results:Mean pre-operative K-angle was 28o. Average loss of correction at final follow up was 3o.Mean correction of K-angle was 140.Moderate to severe loss of correction of K- angle was observed in 4 patients. Mild to moderate pain in 5 patients treated with analgesics. Average TSM subsidence was 3mm.Conclusions:Bone graft composite provides stable biomechanical support to deficient anterior column in burst fractures and allows early rehabilitation and mobilization. Neural recovery may occur after anterior decompression, stabilization and fusion with TSM-Bone graft composite in dorsolumbar burst fractures with incomplete cord injury.

The arithmetic mean and the non-linear median operation, provide two slightly different abstractions of any discrete dataset. The differential information contained in these two filtering operations reveal several trends in the processed data. This principle has been applied towards the process of deriving robust binary PINs from gray-scale images by first extracting pseudo-images by random rectangular downsampling of a parent image, fusing them with a median operation and then mean thresholding the filtered set. Robustness against recompression, noise and most importantly geometric operations, stems from two factors: (a) the random sampling which provides elasticity and (b) the process of selecting the PIN-thresholds as the arithmetic mean filtered set.

In this paper, a hybrid fingerprint matching algorithm based on phase correlation and minutiae points is proposed to improve the matching accuracy in fingerprint verification system. The phase correlation function which operates in the Fourier domain is used to determine the rotational offset while the core point detection algorithm is used to provide translational adjustment. Finally to match the aligned input and template minutiae sets, a Weighted Average Distance (WAD) measure is introduced. Experimental results show that the proposed method has an Equal Error Rate of 1.02% and 0.77% on two public domain databases FVC2002-DB1a and FVC2002-DB2a respectively.

Conventional traitor tracing approaches focus on creating associations between multimedia fingerprints through anti-collusion codes. In this paper, we explore the possibility of creating a code and embedding domain independent fingerprint detection strategy by performing an eigen-decomposition of the fingerprinted images. The resultant Basis which spans the space of valid fingerprinted images and their respective projections are used as side-information for traitor tracing. Any colluded image is projected onto this eigen-space and the nearest neighbours are identified as potential traitors. The proposed architecture was found to be sensitive to linear and non-linear collusion operations permitting traceability, but remained robust to filtering, JPEG compression and noise addition.

Traditionally message authentication codes (MACs) and the process of encryption are treated as orthogonal security mechanisms, where MACs are used to prove data authenticity while encryption is used to preserve confidentiality. In this paper we propose to integrate the two in a framework where the parameterized hash value of an encrypted image is designed to be the same as the hash value of the parent unencrypted original image. This form of privacy preserving authentication has applications in database searches containing sensitive files such as sets of patient's X-ray records. The main challenge here is to develop a parametric hashing algorithm that is invariant to encryption by allowing a small part of the statistical signature of the original image to emerge despite the encryption process. Since the hash value is computed without decrypting the original data, one can prove authenticity without actually revealing the information. Our contribution is twofold: (1) Formulation of this new problem of authenticating encrypted data and (2) Development of a simple hashing algorithm applicable to encrypted images.

In perfect secret sharing schemes, invalid coalitions of shares do not reveal any information regarding the secret. However, in secret sharing constructions which are non-perfect in nature, every illegitimate coalition leaks out information not just pertaining to the parent secret but also related to other shares in circulation. In this paper the information flow and leakage through shares generated by a non-perfect scheme called MIX-SPLIT is examined with the help of a geometric model, developed with a 3-out-of-3 sharing example in view. Finally the principle of controlled leakage has been applied towards a secure biometric-based group authentication system. No fingerprint template storage is required and access to the respective user biometric PINs are made completely secure.

Background: The tibial plateau fracture is a typical example of complex periarticular trauma, which has been and continues to be a considerable clinical challenge to the orthopaedician in terms of decision-making in management. Residual functional deficits in case of complex knee trauma treated with open reduction and internal fixation even when delayed till recovery of soft tissues were found to be significant. With this knowledge, we proposed to conduct a study on the effectiveness of the Illizarov technique in management of displaced tibial plateau fractures. Methods: Prospective study of twenty two cases of complex tibial plateau fracture treated with Illizarov fixator done in Sri Ramachandra medical centre between June 2011-June 2013. The inclusion criteria were closed Schatzker type V & type VI tibial plateau fracture. The exclusion criteria were Schatzker Type I, II, III and IV tibial plateau fracture and any other associated lower limb fractures. The post op outcome was measured using the knee society score at each follow up and the final scoring was done at the final follow up. Results: The average knee society knee score was 78.09 and the average function score was 93.5. We had 95% of excellent/good result. Conclusions: Illizarov technique gives good to excellent results in high energy closed and open tibial condylar fractures. The Illizarov method offers significant advantage in the case of fracture blisters with impending compartment syndrome associated with closed high energy tibial plateau fractures. Illizarov technique provides the greatest advantage of ability to maintain the functional status of the extremity.

Wireless Multimedia applications increasingly demand low power and bandwidth efficient techniques for content protection. Existing partial encryption methodologies juggle with the difficulties of trading decoding complexity for a higher bit rate, maintaining format compliance yet minimizing the number of components that need to be encrypted, to produce the desired distortion. Source based fingerprinting schemes for content tracking are neither bandwidth efficient nor computationally efficient when coupled with encryption. Thus inspired by the chameleon cipher we present a new algorithm for fingerprinting JPEG compressed images and a simple analytical model based on joint fingerprinting and decryption, as a computationally and bandwidth efficient solution for content protection and tracking.