Three-dimensional magnetic resonance imaging-based statistical shape analysis and machine learning-based prediction of patellofemoral instability.

This study performed three-dimensional (3D) magnetic resonance imaging (MRI)-based statistical shape analysis (SSA) by comparing patellofemoral instability (PFI) and normal femur models, and developed a machine learning (ML)-based prediction model. Twenty (19 patients) and 31 MRI scans (30 patients) of femurs with PFI and normal femurs, respectively, were used. Bone and cartilage segmentation of the distal femurs was performed and subsequently converted into 3D reconstructed models. The pointwise distance map showed anterior elevation of the trochlea, particularly at the central floor of the proximal trochlea, in the PFI models compared with the normal models. Principal component analysis examined shape variations in the PFI group, and several principal components exhibited shape variations in the trochlear floor and intercondylar width. Multivariate analysis showed that these shape components were significantly correlated with the PFI/non-PFI distinction after adjusting for age and sex. Our ML-based prediction model for PFI achieved a strong predictive performance with an accuracy of 0.909 ± 0.015, and an area under the curve of 0.939 ± 0.009 when using a support vector machine with a linear kernel. This study demonstrated that 3D MRI-based SSA can realistically visualize statistical results on surface models and may facilitate the understanding of complex shape features.

Longitudinal Morphometric Changes in the Corticospinal Tract Shape After Hemorrhagic Stroke.

Deep intracerebral hemorrhage (ICH) exerts a direct force on corticospinal tracts (CST) causing shape deformation. Using serial MRI, Generalized Procrustes Analysis (GPA), and Principal Components Analysis (PCA), we temporally evaluated the change in CST shape. Thirty-five deep ICH patients with ipsilesional-CST deformation were serially imaged on a 3T-MRI with a median imaging time of day-2 and 84 of onset. Anatomical and diffusion tensor images (DTI) were acquired. Using DTI color-coded maps, 15 landmarks were drawn on each CST and the centroids were computed in 3 dimensions. The contralesional-CST landmarks were used as a reference. The GPA outlined the shape coordinates and we superimposed the ipsilesional-CST shape at the two-time points. A multivariate PCA was applied to identify eigenvectors associated with the highest percentile of change. The first three principal components representing CST deformation along the left-right (PC1), anterior-posterior (PC2), and superior-inferior (PC3) respectively were responsible for 57.9% of shape variance. The PC1 (36.1%, p < 0.0001) and PC3 (9.58%, p < 0.01) showed a significant deformation between the two-time points. Compared to the contralesional-CST, the ipsilesional PC scores were significantly (p < 0.0001) different only at the first-timepoint. A significant positive association between the ipsilesional-CST deformation and hematoma volume was observed. We present a novel method to quantify CST deformation caused by ICH. Deformation most often occurs in left-right axis (PC1) and superior-inferior (PC3) directions. As compared to the reference, the significant temporal difference at the first time point suggests CST restoration over time.

Changes in the Sensory Odor Profile during Chorizo Maturation and Their Relationship with Volatile Compound Patterns by Partial Least Square Regression (PLS).

Odor is one of the most important attributes to determine the overall acceptance of a product. The aim of this investigation is to evaluate the changes in the odor profile and the volatile compounds during thirty-three days of ripening to obtain the pattern of volatile compounds necessary to integrate the odor profile of chorizo (fermented sausage), using Partial Least Squares (PLS). The chili and pork meat odors were predominant during the first five days, vinegar and fermented odors at days twelve and nineteen days, and finally a rancid odor predominated at the end. Only the vinegar, rancid, and fermented odors could be predicted with a good fit model, with the R2 coefficient above 0.5, using linear PLS, and the pork meat odor using logarithmic PLS. Each group of volatile compounds interacted in different ways; esters had a positive influence on the vinegar and rancid odors, but a negative on the fermented odor. Some volatile compounds contributed to more than one odor, such as hexanal, ethanol, and ethyl octanoate. This work allowed us to understand the pattern of volatile compounds required to generate some of the specific odors of chorizo; further studies are required to explore the effect of other food components on these patterns of odors.

Sensory descriptors for three edible Chilean seaweeds and their relations to umami components and instrumental texture.

Although seaweeds exhibit many benefits as a food source, few studies have characterized their sensory attributes. An expert nine-member panel developed a vocabulary with 25 descriptors to describe the appearance, aroma, flavor, texture, and aftertaste of raw and cooked seaweeds consumed in Chile: Durvillaea antarctica, Pyropia spp., and Ulva lactuca. Subsequently, the vocabulary was used in a ranking descriptive analysis (RDA) to evaluate the sensory properties and relate them with physicochemical and physical data. Sensory attributes of the three seaweeds were very different from each other but similar between treatments (raw and cooked). Pyropia spp., both cooked and hydrated, had the highest glutamate content (310 and 324 mg (100 g) -1 d.w., respectively), and was perceived by the sensory panel as having the most umami taste. Cooked D. antarctica was perceived as sweeter, had more caramel notes than the hydrated seaweed and was sensed as cartilaginous and hard in accordance with its mechanical properties. Generalized Procrustes analysis revealed that D. antarctica exhibited most of the desirable descriptors, such as caramel, umami and marine aromas while U. lactuca was described as bitter and moldy. This primary vocabulary can assist food scientists and chefs in the development of seaweed products and dishes for the consumer market.

Measurement error using a SeeMaLab structured light 3D scanner against a Microscribe 3D digitizer.

BACKGROUND: Geometric morphometrics is a powerful approach to capture and quantify morphological shape variation. Both 3D digitizer arms and structured light surface scanners are portable, easy to use, and relatively cheap, which makes these two capturing devices obvious choices for geometric morphometrics. While digitizer arms have been the "gold standard", benefits of having full 3D models are manifold. We assessed the measurement error and investigate bias associated with the use of an open-source, high-resolution structured light scanner called SeeMaLab against the popular Microscribe 3D digitizer arm. METHODOLOGY: The analyses were based on 22 grey seal (Halichoerus grypus) skulls. 31 fixed anatomical landmarks were annotated both directly using a Microscribe 3D digitizer and on reconstructed 3D digital models created from structured light surface scans. Each skull was scanned twice. Two operators annotated the landmarks, each twice on all the skulls and 3D models, allowing for the investigation of multiple sources of measurement error. We performed multiple Procrustes ANOVAs to compare the two devices in terms of within- and between-operator error, to quantify the measurement error induced by device, to compare between-device error with other sources of variation, and to assess the level of scanning-related error. We investigated the presence of general shape bias due to device and operator. RESULTS: Similar precision was obtained with both devices. If landmarks that were identified as less clearly defined and thus harder to place were omitted, the scanner pipeline would achieve higher precision than the digitizer. Between-operator error was biased and seemed to be smaller when using the scanner pipeline. There were systematic differences between devices, which was mainly driven by landmarks less clearly defined. The factors device, operator and landmark replica were all statistically significant and of similar size, but were minor sources of total shape variation, compared to the biological variation among grey seal skulls. The scanning-related error was small compared to all other error sources. CONCLUSIONS: As the scanner showed precision similar to the digitizer, a scanner should be used if the advantages of obtaining detailed 3D models of a specimen are desired. To obtain high precision, a pre-study should be conducted to identify difficult landmarks. Due to the observed bias, data from different devices and/or operators should not be combined when the expected biological variation is small, without testing the landmarks for repeatability across platforms and operators. For any study necessitating the combination of landmark measurements from different operators, the scanner pipeline will be better suited. The small scanning-related error indicates that by following the same scanning protocol, different operators can be involved in the scanning process without introducing significant error.

Developing a Descriptive Sensory Characterization of Flour Tortilla Applying Flash Profile.

For any food, it is important to know consumption, preference, and the characteristics as quality parameters that are important to consumers of a product. The descriptive methodologies are an important tool to know the quality attributes of the products. Within these methodologies is the flash profile (FP), which is based on the generation of the distinctive attributes of the products without any expensive and time-consuming training sessions. The aim of this research was to study the consumption and preference of flour tortillas by consumers and to develop the descriptive characterization of the tortillas by using the flash profile method. The wheat flour tortillas used were two commercial and two handcrafted samples. Ten experienced panelists participated as the FP panel. The panelists generated 22 descriptors, six for texture, seven for appearance, five for odor, and four for flavor. These descriptors differentiate the samples of the flour tortillas. The panelists' performance was assessed using the consensus index (Rc = 0.508). The first two dimensions of the Generalized Procrustes Analysis represent 83.78% of the data variability. Flash profile proved to be an easy and rapid technique that allowed the distinctive attributes of flour tortillas to be obtained.

3D revelation of phenotypic variation, evolutionary allometry, and ancestral states of corolla shape: a case study of clade Corytholoma (subtribe Ligeriinae, family Gesneriaceae).

BACKGROUND: Quantification of corolla shape variations helps biologists to investigate plant diversity and evolution. 3D images capture the genuine structure and provide comprehensive spatial information. RESULTS: This study applied X-ray micro-computed tomography (µCT) to acquire 3D structures of the corollas of clade Corytholoma and extracted a set of 415 3D landmarks from each specimen. By applying the geometric morphometrics (GM) to the landmarks, the first 4 principal components (PCs) in the 3D shape and 3D form analyses, respectively, accounted for 87.86% and 96.34% of the total variance. The centroid sizes of the corollas only accounted for 5.46% of the corolla shape variation, suggesting that the evolutionary allometry was weak. The 4 morphological traits corresponding to the 4 shape PCs were defined as tube curvature, lobe area, tube dilation, and lobe recurvation. Tube curvature and tube dilation were strongly associated with the pollination type and contained phylogenetic signals in clade Corytholoma. The landmarks were further used to reconstruct corolla shapes at the ancestral states. CONCLUSIONS: With the integration of µCT imaging into GM, the proposed approach boosted the precision in quantifying corolla traits and improved the understanding of the morphological traits corresponding to the pollination type, impact of size on shape variation, and evolution of corolla shape in clade Corytholoma.

Flash Profile as an effective method for assessment of odor profile in three different fishes.

Fish has a high nutritional value and its regular consumption is associated with lower probabilities of cardiovascular diseases. Brazilians are characterized as low fish consumers, and the fish odor significantly affects the consumer acceptance or rejection. Grass carp, catfish and pacu are three fishes with high potential of farming in the Southwestern region of Paraná, Brazil, and worldwide. The characterization of their odor profile using the Flash Profile technique is the focus of this study. With assessors selected for odor analysis, the assessment of samples of the three species of fish was carried out in two stages (1) generation of attributes using the Grid method and (2) evaluation of the fish samples, according to attributes individually selected, with 9 cm unstructured and anchored scales. The data was processed by using generalized procrustes analysis. A total of 19 selected assessors were able to recognize and describe the stimuli perceived by the different substances. The terminology surveyed revealed the term "fishy" as the most cited, followed by "pond water", "fat" and "rancid". Flash Profile provided a precise odor characterization and discrimination with quantitative differences among the fishes. Pacu was the species related to the terms earth, fat, plant, fridge and wood. Earth, viscera, putrid, fat, burnt oil, blood, fishy odor and acid characterized the catfish, related to the most negative descriptors. Grass carp was associated to the terms pond water, fishy odor, rancidity, fat and grass.

A multilevel exploration of Avena strigosa diversity as a prelude to promote alternative crop.

BACKGROUND: Sand oat (Avena strigosa Schreb.), one of the four cultivated species of the genus Avena, could be considered as another alternative crop. In gene banks 865 germplasm samples of this species have been preserved that have not been thoroughly investigated so far. The results of phenotyping (36 traits), isoenzymatic (12 systems) and genetic (8 pairs of Sequence-related amplified polymorphism markers) variation were used to obtain the complete description of 56 accessions diversity originated from different parts of world. RESULTS: Breeded and weedy forms represented similar pool of morphological traits that indicated a short-term and extensive breeding process, albeit all accessions which we classified as cultivated were characterized by better grain and green mass parameters compared to the weedy ones. Isoenzymes showed relationships with geographical origin, which was not possible to detect by SRAP markers. There was no similarity between morphological and biochemical results. The polymorphism level of SRAP markers was lower than indicated by the available literature data for other species, however it may result from the analysis of pooled samples of accessions with a high internal variability. The extensive type of breeding and its relatively short duration was also reflected in the population structure results. Joint analysis revealed that a secondary centre of diversity is being created in South America and that it has its genealogy from the Iberian Peninsula. CONCLUSIONS: Despite the relatively large representation of this species is in various gene banks, it is highly probable that the vast majority of stored worldwide accessions are duplicates, and the protected gene pool is relatively narrow. Sand oat meets all the requirements for an alternative crop species, but further studies are needed to identify the genotypes/populations with the most favourable distribution of utility and quality parameters.

Morphometric geometrical analysis to determine the centre of the acetabular component placement in Crowe type IV hips undergoing total hip arthroplasty.

AIMS: We analyzed the acetabular morphology of Crowe type IV hips using CT data to identify a landmark for the ideal placement of the centre of the acetabular component, as assessed by morphometric geometrical analysis, and its reliability. PATIENTS AND METHODS: A total of 52 Crowe IV hips (42 patients; seven male, 35 female; mean age 68.5 years (32 to 82)) and 50 normal hips (50 patients; eight male, 42 female; mean age 60.7 years (34 to 86)) undergoing total hip arthroplasty were retrospectively identified. In this CT-based simulation study, the acetabular component was positioned at the true acetabulum with a radiological inclination of 40° and anteversion of 20°. Acetabular shape and the position of the centre of the acetabular component were analyzed by morphometric geometrical analysis using the generalized Procrustes analysis. RESULTS: The acetabular shapes of Crowe IV hips were distinctively triangular; the ideal position of the centre of the acetabular component was superior on the posterior bony wall. The first and second relative warps explained 34.2% and 18.4% of the variance, respectively, compared with that of 28.6% and 18.0% in normal hips. We defined the landmark as one-third the distance from top on the posterior bony wall in Crowe IV hips. The average distance from the centre of the acetabular component was 5.6 mm. CONCLUSION: Crowe IV hips are distinctively triangular; the point one-third from the top on the posterior bony wall was a useful landmark for placing the acetabular component.

Comparison between quantitative descriptive analysis and flash profile in profiling the sensory properties of commercial red sufu (Chinese fermented soybean curd).

BACKGROUND: Red sufu is a type of sufu produced by solid-state fermentation of soybean curd and coloration with red mold rice. The purposes of this study were: (i) to characterize commercial red sufu samples using the quantitative descriptive analysis (QDA) and flash profile (FP) by ten trained and ten untrained panelists, respectively; (ii) to compare the differences in panel performance, descriptive abilities and sensory maps between the two methodologies; and (iii) to compare the efficiency between QDA and FP using red sufu as the matrix. Techniques in multivariate analysis were utilized to explore the data. RESULTS: Results from generalized procrustes analysis (GPA) showed that panel performance by QDA was more repeatable and reached higher homogeneity than that by FP. Despite the confidence ellipse results of the 12 red sufus being better discriminated by QDA, the RV coefficient was high (RV = 0.852) between the configurations of the two-dimensional model (F1 and F2) of the two methodologies, indicating that the two methods are similar and closely related. Overall, QDA provided more accurate and detailed information, while FP provided a similar sensory map on product location and descriptive results. CONCLUSION: The FP technique appeared to be an efficient alternative approach to quickly evaluate sensory properties, including appearance, flavor, aroma and textural properties of an array of red sufu products. © 2018 Society of Chemical Industry.

Pelvic and Lower Gastrointestinal Tract Anatomical Characterization of the Average Male.

OBJECTIVE: Colorectal surgeons report difficulty in positioning surgical devices in males, particularly those with a narrower pelvis. The objectives of this study were to (1) characterize the anatomy of the pelvis and surrounding soft tissue from magnetic resonance and computed tomography scans from 10 average males (175 cm, 78 kg) and (2) develop a model representing the mean configuration to assess variability. METHODS: The anatomy was characterized from existing scans using segmentation and registration techniques. Size and shape variation in the pelvis and soft tissue morphology was characterized using the Generalized Procrustes Analysis to compute the mean configuration. RESULTS: There was considerable variability in volume of the psoas, connective tissue, and pelvis and in surface area of the mesorectum, pelvis, and connective tissue. Subject height was positively correlated with mesorectum surface area (P = .028, R2 = 0.47) and pelvis volume ( P = .041, R2 = 0.43). The anterior-posterior distance between the inferior pelvic floor muscle and pubic symphysis was positively correlated with subject height ( P = .043, r = 0.65). The angle between the superior mesorectum and sacral promontory was negatively correlated with subject height ( P = .042, r = -0.65). The pelvic inlet was positively correlated with subject weight ( P = .001, r = 0.89). CONCLUSIONS: There was considerable variability in organ volume and surface area among average males with some correlations to subject height and weight. A physical trainer model created from these data helped surgeons trial and assess device prototypes in a controllable environment.

Low resolution scans can provide a sufficiently accurate, cost- and time-effective alternative to high resolution scans for 3D shape analyses.

BACKGROUND: Advances in 3D shape capture technology have made powerful shape analyses, such as geometric morphometrics, more feasible. While the highly accurate micro-computed tomography (µCT) scanners have been the "gold standard," recent improvements in 3D surface scanners may make this technology a faster, portable, and cost-effective alternative. Several studies have already compared the two devices but all use relatively large specimens such as human crania. Here we perform shape analyses on Australia's smallest rodent to test whether a 3D scanner produces similar results to a µCT scanner. METHODS: We captured 19 delicate mouse (Pseudomys delicatulus) crania with a µCT scanner and a 3D scanner for geometric morphometrics. We ran multiple Procrustes ANOVAs to test how variation due to scan device compared to other sources such as biologically relevant variation and operator error. We quantified operator error as levels of variation and repeatability. Further, we tested if the two devices performed differently at classifying individuals based on sexual dimorphism. Finally, we inspected scatterplots of principal component analysis (PCA) scores for non-random patterns. RESULTS: In all Procrustes ANOVAs, regardless of factors included, differences between individuals contributed the most to total variation. The PCA plots reflect this in how the individuals are dispersed. Including only the symmetric component of shape increased the biological signal relative to variation due to device and due to error. 3D scans showed a higher level of operator error as evidenced by a greater spread of their replicates on the PCA, a higher level of multivariate variation, and a lower repeatability score. However, the 3D scan and µCT scan datasets performed identically in classifying individuals based on intra-specific patterns of sexual dimorphism. DISCUSSION: Compared to µCT scans, we find that even low resolution 3D scans of very small specimens are sufficiently accurate to classify intra-specific differences. We also make three recommendations for best use of low resolution data. First, we recommend that extreme caution should be taken when analyzing the asymmetric component of shape variation. Second, using 3D scans generates more random error due to increased landmarking difficulty, therefore users should be conservative in landmark choice and avoid multiple operators. Third, using 3D scans introduces a source of systematic error relative to µCT scans, therefore we recommend not combining them when possible, especially in studies expecting little biological variation. Our findings support increased use of low resolution 3D scans for most morphological studies; they are likely also applicable to low resolution scans of large specimens made in a medical CT scanner. As most vertebrates are relatively small, we anticipate our results will bolster more researchers in designing affordable large scale studies on small specimens with 3D surface scanners.

Characterization of the age-dependent shape of the pediatric thoracic spine and vertebrae using generalized procrustes analysis.

Generalized Procrustes Analysis (GPA) is a superimposition method used to generate size-invariant distributions of homologous landmark points. Several studies have used GPA to assess the three-dimensional (3D) shapes of or to evaluate sex-related differences in the human brain, skull, rib cage, pelvis and lower limbs. Previous studies of the pediatric thoracic vertebrae suggest that they may undergo changes in shape asa result of normative growth. This study uses GPA and second order polynomial equations to model growth and age- and sex-related changes in shape of the pediatric thoracic spine. We present a thorough analysis of the normative 3D shape, size, and orientation of the pediatric thoracic spine and vertebrae as well as equations which can be used to generate models of the thoracic spine and vertebrae for any age between 1 and 19years. Such models could be used to create more accurate 3D reconstructions of the thoracic spine, generate improved age-specific geometries for finite element models (FEMs) and used to assist clinicians with patient-specific planning and surgical interventions for spine deformity.

A potential pitfall in studies of biological shape: Does size matter?

The number of published studies using geometric morphometrics (GM) for analysing biological shape has increased steadily since the beginning of the 1990s, covering multiple research areas such as ecology, evolution, development, taxonomy and palaeontology. Unfortunately, we have observed that many published studies using GM do not evaluate the potential allometric effects of size on shape, which normally require consideration or assessment. This might lead to misinterpretations and flawed conclusions in certain cases, especially when size effects explain a large part of the shape variation. We assessed, for the first time and in a systematic manner, how often published studies that have applied GM consider the potential effects of allometry on shape. We reviewed the 300 most recent published papers that used GM for studying biological shape. We also estimated how much of the shape variation was explained by allometric effects in the reviewed papers. More than one-third (38%) of the reviewed studies did not consider the allometric component of shape variation. In studies where the allometric component was taken into account, it was significant in 88% of the cases, explaining up to 87.3% of total shape variation. We believe that one reason that may cause the observed results is a misunderstanding of the process that superimposes landmark configurations, i.e. the Generalized Procrustes Analysis, which removes isometric effects of size on shape, but not allometric effects. Allometry can be a crucial component of shape variation. We urge authors to address, and report, size effects in studies of biological shape. However, we do not propose to always remove size effects, but rather to evaluate the research question with and without the allometric component of shape variation. This approach can certainly provide a thorough understanding of how much size contributes to the observed shaped variation.

Digital Morphometrics of Two North American Grapevines (Vitis: Vitaceae) Quantifies Leaf Variation between Species, within Species, and among Individuals.

Recent studies have demonstrated that grapevine (Vitis spp.) leaf shape can be quantified using digital approaches which indicate phylogenetic signal in leaf shape, discernible patterns of developmental context within single leaves, and signatures of local environmental conditions. Here, we extend this work by quantifying intra-individual, intraspecific, and interspecific variation in leaf morphology in accessions of North American Vitis riparia and V. rupestris in a common environment. For each species at least four clonal replicates of multiple genotypes were grown in the Missouri Botanical Garden Kemper Center for Home Gardening. All leaves from a single shoot were harvested and scanned leaf images were used to conduct generalized Procrustes analysis, linear discriminant analysis, and elliptical Fourier analysis. Leaf shapes displayed genotype-specific signatures and species distinctions consistent with taxonomic classifications. Leaf shape variation within genotypes and among clones was the result of pest and pathogen-induced leaf damage that alters leaf morphology. Significant trends in leaf damage caused by disease and infestation were non-random with respect to leaf position on the shoot. Digital morphometrics is a powerful tool for assessing leaf shape variation among species, genotypes, and clones under common conditions and suggests biotic factors such as pests and pathogens as important drivers influencing leaf shape.

Genetic and Morphological Differentiation of the Semiterrestrial Crab Armases angustipes (Brachyura: Sesarmidae) along the Brazilian Coast.

The genetic and morphometric population structures of the semiterrestrial crab Armases angustipes from along the Brazilian coast were examined. The influence of the Central South Equatorial Current on larval dispersal of A. angustipes also was evaluated. Six populations were sampled from estuarine areas in São Luis do Maranhão, Maranhão; Natal, Rio Grande do Norte; Maceió, Alagoas; Ilhéus, Bahia; Aracruz, Espírito Santo; and Guaratuba, Paraná. Patterns of genetic differentiation were assessed using DNA sequence data corresponding to parts of the mitochondrial cytochrome c oxidase subunit 1. Geometric morphometric techniques were used to evaluate morphological variation in shape and size of the carapace and right cheliped propodus. Our results revealed low genetic variability and lack of phylogeographic structure; geometric morphometrics showed statistically significant morphological differentiation and geographic structuring. Our data indicate the absence of possible barriers to gene flow for this mobile species, and no clear correlation of morphological or genetic variation with ocean currents and/or geographic distance. Our results also suggest that historical geological and climatological events and/or possible bottleneck effects influenced the current low genetic variability among the populations of A. angustipes.

Comprehensive Methods for Leaf Geometric Morphometric Analyses.

Leaf morphometrics are used frequently by several disciplines, including taxonomists, systematists, developmental biologists, morphologists, agronomists, and plant breeders to name just a few. Leaf shape is highly variable and can be used for identifying species or genotypes, developmental patterning within and among individuals, assessing plant health, and measuring environmental impacts on plant phenotype. Traditional leaf morphometrics requires hand tools and access to specimens, but modern efforts to digitize botanical collections make digital morphometrics a readily accessible and scientifically rigorous option. Here we provide detailed instructions for performing some of the most informative digital geometric morphometric analyses available: generalized Procrustes analysis, elliptical Fourier analysis, and shape features. This comprehensive procedure for leaf shape analysis is comprised of six main sections: A) scanning of material, B) acquiring landmarks, C) analysis of landmark data, D) isolating leaf outlines, E) analysis of leaf outlines, and F) shape features. This protocol provides a detailed reference for applying landmark and outline analysis to leaf shape as well as describing leaf shape features, thus empowering researchers to perform high throughput phenotyping for diverse applications.

Artificial neural networks and geometric morphometric methods as a means for classification: A case-study using teeth from Carcharhinus sp. (Carcharhinidae).

Over the past few decades, geometric morphometric methods have become increasingly popular and powerful tools to describe morphological data while over the same period artificial neural networks have had a similar rise in the classification of specimens to preconceived groups. However, there has been little research into how well these two systems operate together, particularly in comparison to preexisting techniques. In this study, geometric morphometric data and multilayer perceptrons, a style of artificial neural network, were used to classify shark teeth from the genus Carcharhinus to species. Three datasets of varying size and species differences were used. We compared the performance of this combination with geometric morphometric data in a linear discriminate function analysis, linear measurements in a linear discriminate function analysis, and a preexisting methodology from the literature that incorporates linear measurements and a two-layered discriminate function analysis. Across datasets, geometric morphometric data in a multilayer perceptron tended to yield modest accuracies but accuracies that varied less across species whereas other methods were able to achieve higher accuracies in some species at the expense of lower accuracies in others. Further, the performance of the two-layered discriminate function analysis illustrates that constraining what material is classified can increase the accuracy of a method. Based on this tradeoff, the best methodology will then depend on the scope of the study and the amount of material available. J. Morphol. 278:131-141, 2017. ©© 2016 Wiley Periodicals,Inc.