Local Gradient Orientation Research Articles

Radiomic Deformation and Textural Heterogeneity (R-DepTH) Descriptor to Characterize Tumor Field Effect: Application to Survival Prediction in Glioblastoma.

The concept of tumor field effect implies that cancer is a systemic disease with its impact way beyond the visible tumor confines. For instance, in Glioblastoma (GBM), an aggressive brain tumor, the increase in intracranial pressure due to tumor burden often leads to brain herniation and poor outcomes. Our work is based on the rationale that highly aggressive tumors tend to grow uncontrollably, leading to pronounced biomechanical tissue deformations in the normal parenchyma, which when combined with local morphological differences in the tumor confines on MRI scans, will comprehensively capture tumor field effect. Specifically, we present an integrated MRI-based descriptor, radiomic-Deformation and Textural Heterogeneity (r-DepTH). This descriptor comprises measurements of the subtle perturbations in tissue deformations throughout the surrounding normal parenchyma due to mass effect. This involves non-rigidly aligning the patients’ MRI scans to a healthy atlas via diffeomorphic registration. The resulting inverse mapping is used to obtain the deformation field magnitudes in the normal parenchyma. These measurements are then combined with a 3D texture descriptor, Co-occurrence of Local Anisotropic Gradient Orientations (COLLAGE), which captures the morphological heterogeneity and infiltration within the tumor confines, on MRI scans. In this work, we extensively evaluated r-DepTH for survival risk-stratification on a total of 207 GBM cases from 3 different cohorts (Cohort 1 (n1 = 53), Cohort 2 (n2 = 75), and Cohort 3 (n3 = 79)), where each of these three cohorts was used as a training set for our model separately, and the other two cohorts were used for testing, independently, for each training experiment. When employing Cohort 1 for training, r-DepTH yielded Concordance indices (C-indices) of 0.7 and 0.65, hazard ratios (HR) and Confidence Intervals (CI) of 10 (6 – 19) and 5 (3 – 8) on Cohorts 2 and 3, respectively. Similarly, training on Cohort 2 yielded C-indices of 0.6 and 0.7, HR and CI of 1 (0.7 – 2) and 3 (2 – 5) on Cohorts 1 and 3, respectively. Finally, training on Cohort 3 yielded C-indices of 0.75 and 0.63, HR and CI of 24 (10 – 57) and 12 (6 – 21) on Cohorts 1 and 2, respectively. Our results show that r-DepTH descriptor may serve as a comprehensive and a robust MRI-based prognostic marker of disease aggressiveness and survival in solid tumors.

Role of an auxilliary subunit of voltage-gated Na+ channels in endothelial functions: Is Navβ3, a novel actor of vascular mechanosignaling?

The endothelial cell (EC) is a central actor to maintain vascular tone and homeostasis. Endothelial dysfunction is responsible for the development of many cardiovascular diseases including atherosclerosis. The endothelial function is mainly stimulated by the shear stress generated by blood flow. The physical forces produced by flow are detected by many mechanosensors in EC, which trigger intracellular signaling pathways influencing endothelial physiology. Among these mechanosensors, several ion channels have been identified. Voltage-gated Na+ (Nav) channels have been shown to be mechanosensitive in cardiac and intestinal cells. These ion channels are also expressed in EC and then we aim to investigate their role in EC responses to shear stress. Confluent Human Umbilical Vein EC (HUVEC) and Human Aorta EC (teloHAEC) were exposed to laminar (LSS, 20 dynes/cm 2 ) or oscillatory shear stress (OSS, 20 dynes/cm 2 , 0.5 Hz frequency) using the Ibidi® pump system. Cell orientation was quantified by local gradient orientation method with Fiji software. RTqPCR and western blot were used to characterize Nav channel subunits at both transcriptional and protein levels. Exposition of HUVEC and TeloHAEC to a LSS for 24 hours and 4 days induced cell orientation in the direction of flow and activated the atheroprotective signaling pathway (KLF2/KLF4, eNOS). Interestingly, shear stress deep changes in several Nav channels subunits expression level. Notably, SCN3B expression, encoding the Ig-CAM Navβ3 protein, was increased with a 2.1 fold increase at 24 hours, 3.9 at 4 days and 12.6 at 7 days in HUVEC and 2.9 at 4 days in teloHAEC. Moreover, trans-resveratrol-induced KLF2 expression led to increase SCN3B expression, suggesting that this atheroprotective transcription factor might be involved in SCN3B regulation. Indeed, OSS, corresponding to an atherogenic flow, did not induce HUVEC alignment in the direction of flow neither activation of atheroprotective signaling pathway nor a significant increase of SCN3B expression. Taken together, our results suggest that Navβ3 could be a putative novel actor in the mechanotransduction of flow and thus in vasculoprotection.

Vein pattern recognition based on RGB images using Monte Carlo simulation and ridge tracking.

In forensic investigations, images of evidence can often be obtained from crimes such as child pornography and masked violent riots. However, identifying criminals is usually very difficult and sometimes impossible because these images usually contain skin of body parts, while their faces and other commonly used biometrics are unavailable. Vein patterns are a potential biometric to solve this problem. Traditional systems use near-infrared (NIR) imaging technologies to obtain vein patterns, which cannot be applied to forensic analysis since only RGB images are available. However, veins are unobservable in RGB images. In this paper, a comprehensive scheme including a vein uncovering algorithm, a vein extraction algorithm, and a vein pattern matching algorithm is presented. Based on the Monte Carlo (MC) simulation of light transmission in a skin optical model, physical parameters corresponding to different skin colors are obtained, and vein patterns are uncovered from the parameter distribution images. After preprocessing with cubic convolution and Gabor filtering, vein lines are extracted based on ridge tracking. Local gradient orientation and the geometric direction of veins are utilized to guarantee the correct tracking direction. Hessian-based Frangi filters are adopted to locate potential veins. In the matching step, effective minutiae are extracted to represent the topology of vein patterns. A modified coherent point drift (CPD) algorithm is proposed utilizing coordinates, Gabor energy values, and curvatures of minutiae to match vein patterns. Comprehensive experiments were carried out to evaluate the proposed three algorithms. Experimental results show the superiority of the proposed algorithms to various state-of-the-art methods.

A Convolutional Neural Networks-Based Approach for Texture Directionality Detection.

The perceived texture directionality is an important, not fully explored image characteristic. In many applications texture directionality detection is of fundamental importance. Several approaches have been proposed, such as the fast Fourier-based method. We recently proposed a method based on the interpolated grey-level co-occurrence matrix (iGLCM), robust to image blur and noise but slower than the Fourier-based method. Here we test the applicability of convolutional neural networks (CNNs) to texture directionality detection. To obtain the large amount of training data required, we built a training dataset consisting of synthetic textures with known directionality and varying perturbation levels. Subsequently, we defined and tested shallow and deep CNN architectures. We present the test results focusing on the CNN architectures and their robustness with respect to image perturbations. We identify the best performing CNN architecture, and compare it with the iGLCM, the Fourier and the local gradient orientation methods. We find that the accuracy of CNN is lower, yet comparable to the iGLCM, and it outperforms the other two methods. As expected, the CNN method shows the highest computing speed. Finally, we demonstrate the best performing CNN on real-life images. Visual analysis suggests that the learned patterns generalize to real-life image data. Hence, CNNs represent a promising approach for texture directionality detection, warranting further investigation.

Optical and SAR Image Registration Using Complexity Analysis and Binary Descriptor in Suburban Areas

Optical and synthetic aperture radar (SAR) image registration is a challenging task due to significant geometric and radiometric differences. In particular, the strong scattering phenomenon in SAR images can seriously affect the registration results. Accordingly, to solve the low repeatability of the key points in optical and SAR images, a complexity analysis scheme is proposed. In the first phase, the complexity distribution diagram is calculated by the threshold sliding window in the edge images obtained from the maximum moment of the phase congruency. Then morphological operation and connected regions are used to obtain the high-complexity regions and mask them to avoid the extraction of the interference points. Next to solve the limitations of the local self-similarity (LSS) descriptor in the optical and SAR image registration, such as poor discrimination, expensive computational complexity, and sensitivity to large geometric and radiometric difference, we propose a binary LSS descriptor (BLSS). We replace the correlation surface of the LSS with the local gradient orientation histogram. Furthermore, we construct descriptors based on the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">XY</i> -coordinate system and convert the correlation of the regions to binary descriptors. Finally, the fast sample consensus (FSC) is used to remove false correspondences. The experiments conducted on several optical and SAR image pairs verify the effectiveness of the proposed algorithm.

Noise Resilient Local Gradient Orientation for Content-Based Image Retrieval

In a world of multimedia information, where users seek accurate results against search query and demand relevant multimedia content retrieval, developing an accurate content-based image retrieval (CBIR) system is difficult due to the presence of noise in the image. The performance of the CBIR system is impaired by this noise. To estimate the distance between the query and database images, CBIR systems use image feature representation. The noise or artifacts present within the visual data might confuse the CBIR when retrieving relevant results. Therefore, we propose Noise Resilient Local Gradient Orientation (NRLGO) feature representation that overcomes the noise factor within the visual information and strengthens the CBIR to retrieve accurate and relevant results. The proposed NRLGO consists of three steps: estimation and removal of noise to protect the local visual structure; extraction of color, texture, and local contrast features; and, at the end, generation of microstructure for visual representation. The Manhattan distance between the query image and the database image is used to measure their similarity. The proposed technique was tested using the Corel dataset, which contains 10000 images from 100 different categories. The outcomes of the experiment signify that the proposed NRLGO has higher retrieval performance in comparison with state-of-the-art techniques.

Tumor Habitat-derived Radiomic Features at Pretreatment MRI That Are Prognostic for Progression-free Survival in Glioblastoma Are Associated with Key Morphologic Attributes at Histopathologic Examination: A Feasibility Study.

To identify radiomic features extracted from the tumor habitat on routine MR images that are prognostic for progression-free survival (PFS) and to assess their morphologic basis with corresponding histopathologic attributes in glioblastoma (GBM). In this retrospective study, 156 pretreatment GBM MR images (gadolinium-enhanced T1-weighted, T2-weighted, and fluid-attenuated inversion recovery [FLAIR] images) were curated. Of these 156 images, 122 were used for training (90 from The Cancer Imaging Archive and 32 from the Cleveland Clinic, acquired between December 1, 2011, and May 1, 2018) and 34 were used for validation. The validation set was obtained from the Ivy Glioblastoma Atlas Project database, for which the percentage extent of 11 histologic attributes was available on corresponding histopathologic specimens of the resected tumor. Following expert annotations of the tumor habitat (necrotic core, enhancing tumor, and FLAIR-hyperintense subcompartments), 1008 radiomic descriptors (eg, Haralick texture features, Laws energy features, co-occurrence of local anisotropic gradient orientations [CoLIAGe]) were extracted from the three MRI sequences. The top radiomic features were obtained from each subcompartment in the training set on the basis of their ability to risk-stratify patients according to PFS. These features were then concatenated to create a radiomics risk score (RRS). The RRS was independently validated on a holdout set. In addition, correlations (P < .05) of RRS features were computed, with the percentage extent of the 11 histopathologic attributes, using Spearman correlation analysis. RRS yielded a concordance index of 0.80 on the validation set and constituted radiomic features, including Laws (capture edges, waves, ripple patterns) and CoLIAGe (capture disease heterogeneity) from enhancing tumor and FLAIR hyperintensity. These radiomic features were correlated with histopathologic attributes associated with disease aggressiveness in GBM, particularly tumor infiltration (P = .0044) and hyperplastic blood vessels (P = .0005). Preliminary findings demonstrated significant associations of prognostic radiomic features with disease-specific histologic attributes, with implications for risk-stratifying patients with GBM for personalized treatment decisions. Supplemental material is available for this article. © RSNA, 2020.

A Novel Weber Local Binary Descriptor for Fingerprint Liveness Detection

In recent years, fingerprint authentication systems have been extensively deployed in various applications, including attendance systems, authentications on smartphones, mobile payment authorizations, as well as various safety certifications. However, similar to the other biometric identification technologies, fingerprint recognition is vulnerable to artificial replicas made from cheap materials, such as silicon, gelatin, etc. Thus, it is especially necessary to distinguish whether a given fingerprint is a live or a spoof one prior to such authentication. In order to solve the problems above, a novel local descriptor named Weber local binary descriptor for fingerprint liveness detection (FLD) has been proposed in this paper. The method consists of two components: the local binary differential excitation component that extracts intensity-variance features and the local binary gradient orientation component that extracts orientation features. The co-occurrence probability of the two components is calculated to construct a discriminative feature vector, which is fed into support vector machine (SVM) classifiers. The effectiveness of the proposed method is intuitively analyzed on the image samples and numerically demonstrated by Mahalanobis distance. Experiments are performed on two public databases from FLD competitions from 2011 and 2013. The results have proved that the proposed method obtains the best detection accuracy among the existing image local descriptors in FLD.

Disorder in Pixel-Level Edge Directions on T1WI Is Associated with the Degree of Radiation Necrosis in Primary and Metastatic Brain Tumors: Preliminary Findings.

Co-occurrence of local anisotropic gradient orientations (COLLAGE) is a recently developed radiomic (computer extracted) feature that captures entropy (measures the degree of disorder) in pixel-level edge directions and was previously shown to distinguish predominant cerebral radiation necrosis from recurrent tumor on gadolinium-contrast T1WI. In this work, we sought to investigate whether COLLAGE measurements from posttreatment gadolinium-contrast T1WI could distinguish varying extents of cerebral radiation necrosis and recurrent tumor classes in a lesion across primary and metastatic brain tumors. On a total of 75 gadolinium-contrast T1WI studies obtained from patients with primary and metastatic brain tumors and nasopharyngeal carcinoma, the extent of cerebral radiation necrosis and recurrent tumor in every brain lesion was histopathologically defined by an expert neuropathologist as the following: 1) "pure" cerebral radiation necrosis; 2) "mixed" pathology with coexistence of cerebral radiation necrosis and recurrent tumors; 3) "predominant" (>80%) cerebral radiation necrosis; 4) predominant (>80%) recurrent tumor; and 5) pure tumor. COLLAGE features were extracted from the expert-annotated ROIs on MR imaging. Statistical comparisons of COLLAGE measurements using first-order statistics were performed across pure, mixed, and predominant pathologies of cerebral radiation necrosis and recurrent tumor using the Wilcoxon rank sum test. COLLAGE features exhibited decreased skewness for patients with pure (0.15 ± 0.12) and predominant cerebral radiation necrosis (0.25 ± 0.09) and were statistically significantly different (P < .05) from those in patients with predominant recurrent tumors, which had highly skewed (0.42 ± 0.21) COLLAGE values. COLLAGE values for the mixed pathology studies were found to lie between predominant cerebral radiation necrosis and recurrent tumor categories. With additional independent multisite validation, COLLAGE measurements might enable noninvasive characterization of the degree of recurrent tumor or cerebral radiation necrosis in gadolinium-contrast T1WI of posttreatment lesions.

No-Reference Image Quality Assessment with Local Gradient Orientations

Image processing methods often introduce distortions, which affect the way an image is subjectively perceived by a human observer. To avoid inconvenient subjective tests in cases in which reference images are not available, it is desirable to develop an automatic no-reference image quality assessment (NR-IQA) technique. In this paper, a novel NR-IQA technique is proposed in which the distributions of local gradient orientations in image regions of different sizes are used to characterize an image. To evaluate the objective quality of an image, its luminance and chrominance channels are processed, as well as their high-order derivatives. Finally, statistics of used perceptual features are mapped to subjective scores by the support vector regression (SVR) technique. The extensive experimental evaluation on six popular IQA benchmark datasets reveals that the proposed technique is highly correlated with subjective scores and outperforms related state-of-the-art hand-crafted and deep learning approaches.

Building area extraction from the high spatial resolution remote sensing imagery

An approach to building area extraction from high-resolution remote sensing imagery is proposed based on the local gradient orientation density function (LGODF) and the bimodal density function (BDF). Firstly, the LGODF is calculated by moving the window of 35 × 35 based on the gradient magnitude and the gradient direction of each pixel in the image. Then, the BDF is obtained by multiplying the movable bimodal Gaussian mixture function and LGODF in each window. Finally, peaks with difference of 90° are searched for in the BDF, and the central point of the corresponding windows are determined as building pixels. For the validity of the proposed method, seven representative sub-images from PLEIADES images covering Shenzhen China are selected. Experimental results reveal that the precision achieve 94.08% and the recall up to 96.70%.

Radiomic features from pretreatment biparametric MRI predict prostate cancer biochemical recurrence: Preliminary findings.

Radiomics or computer-extracted texture features derived from MRI have been shown to help quantitatively characterize prostate cancer (PCa). Radiomics have not been explored depth in the context of predicting biochemical recurrence (BCR) of PCa. To identify a set of radiomic features derived from pretreatment biparametric MRI (bpMRI) that may be predictive of PCa BCR. Retrospective. In all, 120 PCa patients from two institutions, I1 and I2 , partitioned into training set D1 (N = 70) from I1 and independent validation set D2 (N = 50) from I2 . All patients were followed for ≥3 years. 3T, T2 -weighted (T2 WI) and apparent diffusion coefficient (ADC) maps derived from diffusion-weighted sequences. PCa regions of interest (ROIs) on T2 WI were annotated by two experienced radiologists. Radiomic features from bpMRI (T2 WI and ADC maps) were extracted from the ROIs. A machine-learning classifier (CBCR ) was trained with the best discriminating set of radiomic features to predict BCR (pBCR ). Wilcoxon rank-sum tests with P < 0.05 were considered statistically significant. Differences in BCR-free survival at 3 years using pBCR was assessed using the Kaplan-Meier method and compared with Gleason Score (GS), PSA, and PIRADS-v2. Distribution statistics of co-occurrence of local anisotropic gradient orientation (CoLlAGe) and Haralick features from T2 WI and ADC were associated with BCR (P < 0.05) on D1 . CBCR predictions resulted in a mean AUC = 0.84 on D1 and AUC = 0.73 on D2 . A significant difference in BCR-free survival between the predicted classes (BCR + and BCR-) was observed (P = 0.02) on D2 compared to those obtained from GS (P = 0.8), PSA (P = 0.93) and PIRADS-v2 (P = 0.23). Radiomic features from pretreatment bpMRI can be predictive of PCa BCR after therapy and may help identify men who would benefit from adjuvant therapy. 4 Technical Efficacy: Stage 5 J. Magn. Reson. Imaging 2018;48:1626-1636.

Neural Network for Nanoscience Scanning Electron Microscope Image Recognition

In this paper we applied transfer learning techniques for image recognition, automatic categorization, and labeling of nanoscience images obtained by scanning electron microscope (SEM). Roughly 20,000 SEM images were manually classified into 10 categories to form a labeled training set, which can be used as a reference set for future applications of deep learning enhanced algorithms in the nanoscience domain. The categories chosen spanned the range of 0-Dimensional (0D) objects such as particles, 1D nanowires and fibres, 2D films and coated surfaces, and 3D patterned surfaces such as pillars. The training set was used to retrain on the SEM dataset and to compare many convolutional neural network models (Inception-v3, Inception-v4, ResNet). We obtained compatible results by performing a feature extraction of the different models on the same dataset. We performed additional analysis of the classifier on a second test set to further investigate the results both on particular cases and from a statistical point of view. Our algorithm was able to successfully classify around 90% of a test dataset consisting of SEM images, while reduced accuracy was found in the case of images at the boundary between two categories or containing elements of multiple categories. In these cases, the image classification did not identify a predominant category with a high score. We used the statistical outcomes from testing to deploy a semi-automatic workflow able to classify and label images generated by the SEM. Finally, a separate training was performed to determine the volume fraction of coherently aligned nanowires in SEM images. The results were compared with what was obtained using the Local Gradient Orientation method. This example demonstrates the versatility and the potential of transfer learning to address specific tasks of interest in nanoscience applications.

Anti-spoofing enabled face recognition based on aggregated local weighted gradient orientation

Spoofing attack is a catastrophic threat for biometric authentication systems. Inspired by the concept of depth map estimation, a novel anti-spoofing technique based on aggregated local weighted gradient orientation (ALWGO) is proposed. We first estimate the depth of the specimen face image. In the next step, highly discriminant ALWGO features are extracted from the depth map. Finally, a sparse representation classifier is trained to distinguish between the genuine and fake faces. This paper particularly addresses the potential of texture gradient features and their variations, on three types of attacks, viz. printed high-definition photographs, warped photographs and videos displayed on mobile phones. The usage of ALWGO features has been extended for further face recognition. Our proposed approach is robust and nonintrusive as compared to many existing methods. Extensive experimental analysis on publicly available databases clearly demonstrates the superiority of our approach for both face spoofing detection and recognition systems.

Automatic detection of urban area from the remote sensing imagery based on improved D-S evidence theory

ABSTRACTUrban area detection from the remote sensing imagery has important significance in dynamic monitoring on the use types of city land and the city development planning. An approach for full-automatic detection of urban area is proposed, which can fuse any number of features. The combination formula of the traditional D-S evidence theory (TDSET) is improved by adding a processing factor for conflicting evidences based on the original fusion function. For the validity of the proposed method, three positive evidences (gradient mean, Harris feature points and spectral homogeneity) as well as one negative evidence (local gradient orientation density) are selected to test three GeoEye images. Experimental results show that the precision of the proposed method achieves 96.89% and its recall is 86.52%, where the recall increases about 50% comparing with the TDSET.

Building-Based Damage Detection From Postquake Image Using Multiple-Feature Analysis

Damaged building detection from high spatial resolution remote sensing image helps to rapid disaster losses assessment. However, the majority of traditional methods relies on only a single category feature of the damaged building. This letter presents a new strategy for detecting damaged buildings from postquake remote sensing image by multiple-feature analysis, in which the integrity of the building edge and the interior roof was both considered. The intactness of the building edge was assessed by proposing a new feature parameter, edge significance (ES), ES using significance test to quantify the difference between the gradient values on the edge and in the edge buffer. In addition, the gradient orientation inside the building was analyzed and local gradient orientation entropy (LOE) parameter was adopted to determine whether the interior roof was damaged. In general, damaged buildings have lower ES values because of broken edges and higher LOE values owing to debris, final decision was made on the basis of both feature parameters. A Quickbird image of Yushu, China, was used in the experiment and, among a total of 327 buildings, 266 were detected correctly. The overall accuracy was 84.10%, which is better than traditional methods.

Kernel Embedding Multiorientation Local Pattern for Image Representation.

Local feature descriptor plays a key role in different image classification applications. Some of these methods such as local binary pattern and image gradient orientations have been proven effective to some extent. However, such traditional descriptors which only utilize single-type features, are deficient to capture the edges and orientations information and intrinsic structure information of images. In this paper, we propose a kernel embedding multiorientation local pattern (MOLP) to address this problem. For a given image, it is first transformed by gradient operators in local regions, which generate multiorientation gradient images containing edges and orientations information of different directions. Then the histogram feature which takes into account the sign component and magnitude component, is extracted to form the refined feature from each orientation gradient image. The refined feature captures more information of the intrinsic structure, and is effective for image representation and classification. Finally, the multiorientation refined features are automatically fused in the kernel embedding discriminant subspace learning model. The extensive experiments on various image classification tasks, such as face recognition, texture classification, object categorization, and palmprint recognition show that MOLP could achieve competitive performance with those state-of-the art methods.

Multi-scale non-local denoising method in neuroimaging.

Non-local means algorithm can remove image noise in a unique way that is contrary to traditional techniques. This is because it not only smooths the image but it also preserves the information details of the image. However, this method suffers from high computational complexity. We propose a multi-scale non-local means method in which adaptive multi-scale technique is implemented. In practice, based on each selected scale, the input image is divided into small blocks. Then, we remove the noise in the given pixel by using only one block. This can overcome the low efficiency problem caused by the original non-local means method. Our proposed method also benefits from the local average gradient orientation. In order to perform evaluation, we compared the processed images based on our technique with the ones by the original and the improved non-local means denoising method. Extensive experiments are conducted and results shows that our method is faster than the original and the improved non-local means method. It is also proven that our implemented method is robust enough to remove noise in the application of neuroimaging.

Improving cross-modal face recognition using polarimetric imaging.

We investigate the performance of polarimetric imaging in the long-wave infrared (LWIR) spectrum for cross-modal face recognition. For this work, polarimetric imagery is generated as stacks of three components: the conventional thermal intensity image (referred to as S0), and the two Stokes images, S1 and S2, which contain combinations of different polarizations. The proposed face recognition algorithm extracts and combines local gradient magnitude and orientation information from S0, S1, and S2 to generate a robust feature set that is well-suited for cross-modal face recognition. Initial results show that polarimetric LWIR-to-visible face recognition achieves an 18% increase in Rank-1 identification rate compared to conventional LWIR-to-visible face recognition. We conclude that a substantial improvement in automatic face recognition performance can be achieved by exploiting the polarization-state of radiance, as compared to using conventional thermal imagery.

Co-occurrence of local anisotropic gradient orientations (CoLIAGe): distinguishing tumor confounders and molecular subtypes on MRI.

We introduce a novel biologically inspired feature descriptor, Co-occurrence of Local Anisotropic Gradient Orientations (CoLlAGe), that captures higher order co-occurrence patterns of local gradient tensors at a pixel level to distinguish disease phenotypes that have similar morphologic appearances. A number of pathologies (e.g. subtypes of breast cancer) have different histologic phenotypes but similar radiographic appearances. While texture features have been previously employed for distinguishing subtly different pathologies, they attempt to capture differences in global intensity patterns. In this paper we attempt to model CoLlAGe to identify higher order co-occurrence patterns of gradient tensors at a pixel level. The assumption behind this new feature is that different pathologies, even though they may have very similar overall texture and appearance on imaging, at a local scale, will have different co-occurring patterns with respect to gradient orientations. We demonstrate the utility of CoLIAGe in distinguishing two subtly different types of pathologies on MRI in the context of brain tumors and breast cancer. In the first problem, we look at CoLlAGe for distinguishing radiation effects from recurrent brain tumors over a cohort of 40 studies, and in the second, discriminating different molecular subtypes of breast cancer over a cohort of 73 studies. For both these challenging cohorts, CoLlAGe was found to have significantly improved classification performance, as compared to the traditional texture features such as Haralick, Gabor, local binary patterns, and histogram of gradients.