Search

Symmetric Positive Definite (SPD) matrices emerge as data descriptors in several applications of computer vision such as object tracking, texture recognition, and diffusion tensor imaging. Clustering these data matrices forms an integral part of these applications, for which soft-clustering algorithms (K-Means, expectation maximization, etc.) are generally used. As is well-known, these algorithms need the number of clusters to be specified, which is difficult when the dataset scales. To address this issue, we resort to the classical nonparametric Bayesian framework by modeling the data as a mixture model using the Dirichlet process (DP) prior. Since these matrices do not conform to the Euclidean geometry, rather belongs to a curved Riemannian manifold,existing DP models cannot be directly applied. Thus, in this paper, we propose a novel DP mixture model framework for SPD matrices. Using the log-determinant divergence as the underlying dissimilarity measure to compare these matrices, and further using the connection between this measure and the Wishart distribution, we derive a novel DPM model based on the Wishart-Inverse-Wishart conjugate pair. We apply this model to several applications in computer vision. Our experiments demonstrate that our model is scalable to the dataset size and at the same time achieves superior accuracy compared to several state-of-the-art parametric and nonparametric clustering algorithms. [ABSTRACT FROM PUBLISHER]

• Perform automatic extraction of phenotypic traits for individual plants on a 3D model. • Point cloud processing algorithms take advantage of geometry and topology of crop. • Compute LAI, stand count, height, leaf length, inter-nodal distance, and leaf angle. • Experiments on artificial and real in-field corn plants. High resolution RGB imagery collected using a UAV and a handheld camera was used with structure from motion to reconstruct 3D canopies of small groups of corn plants. A methodology for the automated extraction of phenotypic characteristics of individual plants is presented based on these 3D reconstructed canopies. Such information can enhance the evaluation of crop traits and provide accurate and frequent statistics for in-season assessment of their changes with growth stage. Industries that target yield optimization and crop hybrid production can benefit greatly from this approach. The use of 3D models provides elevated information content, when compared to alternative planar methods, mainly due to the alleviation of leaf occlusions. High resolution images of corn stalks are collected and used to obtain 3D models for individual plants. Based on those extracted 3D point clouds, the calculation of phenotypic characteristics are obtained, such as the number of plants in an area, the Leaf Area Index (LAI), the individual and average plant height, the individual leaf length, the location and the angles of leaves with respect to the stem. An experimental validation using both artificial corn plants emulating real world scenarios and real corn plants in different growth stages, supports the accuracy of the proposed methodology. Our experiments conclude that phenotypic characteristics of individual plants can be extracted automatically with high accuracy based on a 3D model. The results include the individual plant segmentation and counting from a given 3D reconstructed field scene with 88.1% accuracy, the Leaf Area Index (LAI) estimation with 92.5% accuracy, the individual plant height computation with 89.2% accuracy, the leaf length extraction with 74.8% accuracy, the measurement of angles between leaves and stems, and the distance between the leaves of the same plant. We interpret the last two variables qualitatively to show that the method can show the trend of the angles to change with respect to the leaf position on the stem as the crops grow. [ABSTRACT FROM AUTHOR]

Representations in the form of Symmetric Positive Definite (SPD) matrices have been popularized in a variety of visual learning applications due to their demonstrated ability to capture rich second-order statistics of visual data. There exist several similarity measures for comparing SPD matrices with documented benefits. However, selecting an appropriate measure for a given problem remains a challenge and in most cases, is the result of a trial-and-error process. In this paper, we propose to learn similarity measures in a data-driven manner. To this end, we capitalize on the $\alpha \beta$ α β -log-det divergence, which is a meta-divergence parametrized by scalars $\alpha$ α and $\beta$ β , subsuming a wide family of popular information divergences on SPD matrices for distinct and discrete values of these parameters. Our key idea is to cast these parameters in a continuum and learn them from data. We systematically extend this idea to learn vector-valued parameters, thereby increasing the expressiveness of the underlying non-linear measure. We conjoin the divergence learning problem with several standard tasks in machine learning, including supervised discriminative dictionary learning and unsupervised SPD matrix clustering. We present Riemannian gradient descent schemes for optimizing our formulations efficiently, and show the usefulness of our method on eight standard computer vision tasks. [ABSTRACT FROM AUTHOR]

In recent years, there has been extensive research on sparse representation of vector-valued signals. In the matrix case, the data points are merely vectorized and treated as vectors thereafter (for example, image patches). However, this approach cannot be used for all matrices, as it may destroy the inherent structure of the data. Symmetric positive definite (SPD) matrices constitute one such class of signals, where their implicit structure of positive eigenvalues is lost upon vectorization. This paper proposes a novel sparse coding technique for positive definite matrices, which respects the structure of the Riemannian manifold and preserves the positivity of their eigenvalues, without resorting to vectorization. Synthetic and real-world computer vision experiments with region covariance descriptors demonstrate the need for and the applicability of the new sparse coding model. This work serves to bridge the gap between the sparse modeling paradigm and the space of positive definite matrices. [ABSTRACT FROM AUTHOR]

A framework for robust foreground detection that works under difficult conditions such as dynamic background and moderately moving camera is presented in this paper. The proposed method includes two main components: coarse scene representation as the union of pixel layers, and foreground detection in video by propagating these layers using a maximum-likelihood assignment. We first cluster into "layers" those pixels that share similar statistics. The entire scene is then modeled as the union of such nonparametric layer-models. An incoming pixel is detected as foreground if it does not adhere to these adaptive models of the background. A principled way of computing thresholds is used to achieve robust detection performance with a prespecified number of false alarms. Correlation between pixels in the spatial vicinity is exploited to deal with camera motion without precise registration or optical flow. The proposed technique adapts to changes in the scene, and allows to automatically convert persistent foreground objects to background and reconvert them to foreground when they become interesting. This simple framework addresses the important problem of robust foreground and unusual region detection, at about 10 frames per second on a standard laptop computer. The presentation of the proposed approach is complemented by results on challenging real data and comparisons with other standard techniques. [ABSTRACT FROM AUTHOR]

A major component of an efficient farming strategy is the precise detection and characterization of plant deficiencies followed by the proper deployment of fertilizers. Through the thoughtful utilization of modern computer vision techniques, it is possible to achieve positive financial and environmental results for these tasks. This work introduces an automation framework that attempts to address the three main drawbacks of existing approaches: 1) lack of generality (methods are tuned for specific data sets); 2) difficulty to apply in variable field conditions; and 3) lack of tool sophistication that limits their applicability. The cultivation of corn lies in the core of the American and global economy with 81.7 million acres harvested only in the USA for the year 2018. The ubiquity of its cultivation makes it an ideal candidate to highlight the large economic benefits from even a small improvement in nutrient deficiency detection. The proposed methodology utilizes drone collected images to detect nitrogen (N) deficiencies in maize fields and assess their severity using low-cost RGB sensors. The proposed methodology is twofold. A low complexity recommendation scheme identifies candidate plants exhibiting N deficiency and, with minimal interaction, assists the annotator in the creation of a training data set that is then used to train an object detection deep neural network. Results on data from experimental fields support the merits of the proposed methodology with mean average precision for the detection of N-deficient leaves reaching 82.3%. Note to Practitioners—The motivation behind this article is the problem of inefficient fertilizer application in corn fields throughout the cultivation season. Current widely spread techniques to counter plant malnutrition suggest the application of excessive amounts of nitrogen fertilizer prior to seeding or the uniform application during the plant growth. These practices result in financial losses and have severe environmental consequences, e.g., the dead zone in the Gulf of Mexico. We propose an automation framework that automatically detects corn nitrogen deficiencies in the field during the plants’ growth, and to achieve our goal, we employ low-cost robotic platforms and RGB sensors. The framework that we developed is able to detect the characteristic pattern of nitrogen deficiency on corn leaves and provide an estimation of the in-field spatial variability of the deficiency. [ABSTRACT FROM AUTHOR]

Mental health disorders are a leading cause of disability in North America. An important aspect in treating mental disorders is early intervention, which dramatically increases the probability of positive outcomes; however, early intervention hinges upon knowledge and detection of risk markers for particular disorders. Ideally, the screening of these risk markers should occur in a community setting, but this is time-consuming and resource-intensive. Assistive systems could greatly aid in the detection of risk markers in a hectic environment like a preschool classroom. This paper presents a multi-sensor system consisting of 5 RGB-D sensors that detects and tracks the location of occupants in a preschool classroom and computes a measure of activity level and proximity between individuals, an index of social functioning. This assistive system operates in near real-time and is able to track occupants and deal with difficult situations both with occupants (children sitting and laying on the ground, hugging, playing dress-up, etc) and their environment (i.e., changing light levels from artificial and natural sources). The system is installed at, and validated on recordings taken from, the Shirley G. Moore Lab School, a research preschool classroom at the University of Minnesota. The work described herein provides the initial groundwork for monitoring basic elements of child behavior; future efforts will be geared toward identifying and tracking more sophisticated behavioral signatures relevant to mental health. [ABSTRACT FROM AUTHOR]

Objectives: The clinical presentation of pediatric obsessive-compulsive disorder (OCD) is heterogeneous, which is a stumbling block to understanding pathophysiology and to developing new treatments. A major shift in psychiatry, embodied in the Research Domain Criteria (RDoC) initiative of National Institute of Mental Health, recognizes the pitfalls of categorizing mental illnesses using diagnostic criteria. Instead, RDoC encourages researchers to use a dimensional approach, focusing on narrower domains of psychopathology to characterize brain-behavior relationships. Our aim in this multidisciplinary pilot study was to use computer vision tools to record OCD behaviors and to cross-validate these behavioral markers with standard clinical measures.Methods: Eighteen youths with OCD and 21 healthy controls completed tasks in an innovation laboratory (free arrangement of objects, hand washing, arrangement of objects on contrasting carpets). Tasks were video-recorded. Videos were coded by blind raters for OCD-related behaviors. Children's Yale-Brown Obsessive Compulsive Scale (CY-BOCS) and other scales were administered. We compared video-recorded measures of behavior in OCD versus healthy controls and correlated video measures and clinical measures of OCD.Results: Behavioral measures on the videos were significantly correlated with specific CY-BOCS dimension scores. During the free arrangement task, more time spent ordering objects and more moves of objects were both significantly associated with higher CY-BOCS ordering/repeating dimension scores. Longer duration of hand washing was significantly correlated with higher scores on CY-BOCS ordering/repeating and forbidden thoughts dimensions. During arrangement of objects on contrasting carpets, more moves and more adjustment of objects were significantly associated with higher CY-BOCS ordering/repeating dimension scores.Conclusion: Preliminary data suggest that measurement of behavior using video recording is a valid approach for quantifying OCD psychopathology. This methodology could serve as a new tool for investigating OCD using an RDoC approach. This objective, novel behavioral measurement technique may benefit both researchers and clinicians in assessing pediatric OCD and in identifying new behavioral markers of OCD. Clinical Trial Registry: Development of an Instrument That Monitors Behaviors Associated With OCD. NCT02866422. http://clinicaltrials.gov. [ABSTRACT FROM AUTHOR]

The early detection of developmental disorders is key to child outcome, allowing interventions to be initiated which promote development and improve prognosis. Research on autism spectrum disorder (ASD) suggests that behavioral signs can be observed late in the first year of life. Many of these studies involve extensive frame-by-frame video observation and analysis of a child's natural behavior. Although nonintrusive, these methods are extremely time-intensive and require a high level of observer training; thus, they are burdensome for clinical and large population research purposes. This work is a first milestone in a long-term project on non-invasive early observation of children in order to aid in risk detection and research of neurodevelopmental disorders. We focus on providing low-cost computer vision tools to measure and identify ASD behavioral signs based on components of the Autism Observation Scale for Infants (AOSI). In particular, we develop algorithms to measure responses to general ASD risk assessment tasks and activities outlined by the AOSI which assess visual attention by tracking facial features. We show results, including comparisons with expert and nonexpert clinicians, which demonstrate that the proposed computer vision tools can capture critical behavioral observations and potentially augment the clinician's behavioral observations obtained fromreal in-clinic assessments. [ABSTRACT FROM AUTHOR]

The abstract of the article "Exploring Visual Design Preferences in Youths With Obsessive-Compulsive Disorder" by Elizabeth C. Harris and colleagues is presented.