Andrew R. Moldenke

Automated insect identification through concatenated histograms of local appearance features: feature vector generation and region detection for deformable objects

This paper describes a computer vision approach to automated rapid-throughput taxonomic identification of stonefly larvae. The long-term objective of this research is to develop a cost-effective method for environmental monitoring based on automated identification of indicator species. Recognition of stonefly larvae is challenging because they are highly articulated, they exhibit a high degree of intraspecies variation in size and color, and some species are difficult to distinguish visually, despite prominent dorsal patterning. The stoneflies are imaged via an apparatus that manipulates the specimens into the field of view of a microscope so that images are obtained under highly repeatable conditions. The images are then classified through a process that involves (a) identification of regions of interest, (b) representation of those regions as SIFT vectors (Lowe, in Int J Comput Vis 60(2):91–110, 2004) (c) classification of the SIFT vectors into learned “features” to form a histogram of detected features, and (d) classification of the feature histogram via state-of-the-art ensemble classification algorithms. The steps (a) to (c) compose the concatenated feature histogram (CFH) method. We apply three region detectors for part (a) above, including a newly developed principal curvature-based region (PCBR) detector. This detector finds stable regions of high curvature via a watershed segmentation algorithm. We compute a separate dictionary of learned features for each region detector, and then concatenate the histograms prior to the final classification step. We evaluate this classification methodology on a task of discriminating among four stonefly taxa, two of which, Calineuria and Doroneuria, are difficult even for experts to discriminate. The results show that the combination of all three detectors gives four-class accuracy of 82% and three-class accuracy (pooling Calineuria and Doro-neuria) of 95%. Each region detector makes a valuable contribution. In particular, our new PCBR detector is able to discriminate Calineuria and Doroneuria much better than the other detectors.

Dictionary-free categorization of very similar objects via stacked evidence trees

Current work in object categorization discriminates among objects that typically possess gross differences which are readily apparent. However, many applications require making much finer distinctions. We address an insect categorization problem that is so challenging that even trained human experts cannot readily categorize images of insects considered in this paper. The state of the art that uses visual dictionaries, when applied to this problem, yields mediocre results (16.1% error). Three possible explanations for this are (a) the dictionaries are unsupervised, (b) the dictionaries lose the detailed information contained in each keypoint, and (c) these methods rely on hand-engineered decisions about dictionary size. This paper presents a novel, dictionary-free methodology. A random forest of trees is first trained to predict the class of an image based on individual keypoint descriptors. A unique aspect of these trees is that they do not make decisions but instead merely record evidence-i.e., the number of descriptors from training examples of each category that reached each leaf of the tree. We provide a mathematical model showing that voting evidence is better than voting decisions. To categorize a new image, descriptors for all detected keypoints are “dropped” through the trees, and the evidence at each leaf is summed to obtain an overall evidence vector. This is then sent to a second-level classifier to make the categorization decision. We achieve excellent performance (6.4% error) on the 9-class STONEFLY9 data set. Also, our method achieves an average AUC of 0.921 on the PASCAL06 VOC, which places it fifth out of 21 methods reported in the literature and demonstrates that the method also works well for generic object categorization.

Insect Production from Temporary and Perennially Flowing Headwater Streams in Western Oregon

ABSTRACT To evaluate the contribution of headwater streams to the forest ecosystem of the Pacific Northwest, we used emergence traps to examine the effect of stream flow (perennial vs. dry-season temporary) on emergent aquatic insect fauna at three sites in the conifer forests of western Oregon. Total density and biomass of aquatic insects were higher in temporary streams than in perennial streams. Taxonomic richness was consistently higher in perennial streams. The number of Chironomidae exceeded those of all other taxa during the spring, but the chironomids were largely replaced by Mycetophilidae as the most abundant taxon during the summer, especially in temporary streams. Trichoptera and Ephemeroptera emerged in higher numbers from perennial than from temporary streams. These results are consistent with our hypothesis that the absence of vertebrate predators (fish and giant salamanders) allows the populations of arthropods in temporary streams to flourish, serving as: (1) a potential source of colonization and (2) an important role in the terrestrial food web as an abundant food source for terrestrial insectivores.

Automated processing and identification of benthic invertebrate samples

Abstract We present a visually based method for the taxonomic identification of benthic invertebrates that automates image capture, image processing, and specimen classification. The BugID system automatically positions and images specimens with minimal user input. Images are then processed with interest operators (machine-learning algorithms for locating informative visual regions) to identify informative pattern features, and this information is used to train a classifier algorithm. Naïve Bayes modeling of stacked decision trees is used to determine whether a specimen is an unknown distractor (taxon not in the training data set) or one of the species in the training set. When tested on images from 9 larval stonefly taxa, BugID correctly identified 94.5% of images, even though small or damaged specimens were included in testing. When distractor taxa (10 common invertebrates not present in the training set) were included to make classification more challenging, overall accuracy decreased but generally was close to 90%. At the equal error rate (EER), 89.5% of stonefly images were correctly classified and the accuracy of nonrejected stoneflies increased to 96.4%, a result suggesting that many difficult-to-identify or poorly imaged stonefly specimens had been rejected prior to classification. BugID is the first system of its kind that allows users to select thresholds for rejection depending on the required use. Rejected images of distractor taxa or difficult specimens can be identified later by a taxonomic expert, and new taxa ultimately can be incorporated into the training set of known taxa. BugID has several advantages over other automated insect classification systems, including automated handling of specimens, the ability to isolate nontarget and novel species, and the ability to identify specimens across different stages of larval development.

Effects of mowing frequency on densities of natural enemies in three Pacific Northwest pear orchards

Effects of mowing frequency on ground cover composition and on numbers of predators, parasitoids, and select phytophagous arthropods in the ground cover of three reduced‐insecticide pear orchards were determined. Concurrent samples taken in the tree canopy (with beating trays) and in the herbicide strips on the orchard floor (with pitfall traps) tested whether counts of natural enemies in these two habitats were also affected by mowing regime. A reduction in frequency of mowing from two to three times per month (= control) to once per month or once per growing season led to increased cover of grasses, broadleaf plants, and broadleaf plants in flower. Sweep net samples of natural enemies in the ground cover were dominated numerically by spiders (Araneae), parasitic Hymenoptera, and predatory Heteroptera, with lesser numbers of other taxa (Syrphidae, Neuroptera, Coccinellidae). Predators and parasitoids showed substantial increases in numbers associated with decreased mowing frequency. Sweep net counts of aphids, Lygus spp. (Heteroptera: Miridae), and leafhoppers/planthoppers, all potential prey of predators, also increased significantly with decreased mowing frequency. In the pitfall samples, only the European earwig (Forficula auricularia L.) (Dermaptera: Forficulidae) exhibited a change in counts associated with mowing treatment; numbers of earwigs in pitfall traps declined as mowing frequency decreased. For the beat tray samples, mean tray counts for most natural enemy taxa were higher in the less frequently mowed plots, but significantly (P < 0.05) so only for two taxa: spiders and a predatory mirid, Deraeocoris brevis (Uhler) (Heteroptera: Miridae). It remains to be determined whether biological control of pests in the tree canopy can be enhanced by manipulating mowing frequency. Questions raised by this study include whether there is extensive movement by natural enemies between the ground cover and tree canopy, and whether plot size affects the likelihood of showing that mowing frequency influences predator densities in the tree canopy.

Response of Ground-Dwelling Arthropods to Different Thinning Intensities in Young Douglas Fir Forests of Western Oregon

Abstract We evaluated the effect of four different forest management techniques, unthinned control and three thinning intensities (light, light with gaps, and heavy thin), on arthropod abundance, diversity, and community structure as an indicator of ecological processes affecting other forest fauna. Ground-dwelling arthropods were collected during 2000–2001, with pitfall traps in June (warm-wet season) and August (hot-dry season) 5 yr after a thinning treatment in 40- to 60-yr-old Douglas fir [Pseudotsuga menziesii (Mirb.) Franco] trees in the Willamette National Forest. We found arthropod abundance and diversity was higher in heavy and light/gap thinning treatments than the other treatments. Additionally, four groups (ants, spiders, camel-crickets, and millipedes) preferred the more intense thinning treatment areas. The abundance of carabids, the third most abundant group, was higher at the unthinned control than any thinning treatment during the wet season, but not during the dry season. Although the immediate disturbance associated with thinning might be expected to decrease population density of fauna such as ground beetles, we hypothesized that the principal effect of thinning was to increase habitat heterogeneity in these uniform plantations and indirectly increase species richness and abundance of soil-dwellers. Nonmetric multidimensional scaling (NMS) of overall arthropod community composition revealed that both seasonality and thinning were highly significant, resulting in four separate clusters of points, with season dominating thinning. Both variables were correlated with litter moisture. The NMS results indicated that ants preferred heavy thinning intensity. Spiders, carabids, and millipedes were positively associated with litter moisture, and camel-crickets were negatively associated with litter moisture. Overall, our results suggest that some dominant groups of ground-dwelling arthropods are sensitive indicators of environmental change, such as forest thinning.

Automated Insect Identification through Concatenated Histograms of Local Appearance Features

This paper describes a fully automated stone fly-larvae classification system using a local features approach. It compares the three region detectors employed by the system: the Hessian-affine detector, the Kadir entropy detector and a new detector we have developed called the principal curvature based region detector (PCBR). It introduces a concatenated feature histogram (CFH) methodology that uses histograms of local region descriptors as feature vectors for classification and compares the results using this methodology to that of Opelt [Opelt, A, et.al., 2006.] on three stonefly identification tasks. Our results indicate that the PCBR detector outperforms the other two detectors on the most difficult discrimination task and that the use of all three detectors outperforms any other configuration. The CFH methodology also outperforms the Opelt methodology in these tasks

Stacked spatial-pyramid kernel: An object-class recognition method to combine scores from random trees

The combination of local features, complementary feature types, and relative position information has been successfully applied to many object-class recognition tasks. Stacking is a common classification approach that combines the results from multiple classifiers, having the added benefit of allowing each classifier to handle a different feature space. However, the standard stacking method by its own nature discards any spatial information contained in the features, because only the combination of raw classification scores are input to the final classifier. The object-class recognition method proposed in this paper combines different feature types in a new stacking framework that efficiently quantizes input data and boosts classification accuracy, while allowing the use of spatial information. This classification method is applied to the task of automated insect-species identification for biomonitoring purposes. The test data set for this work contains 4722 images with 29 insect species, belonging to the three most common orders used to measure stream water quality, several of which are closely related and very difficult to distinguish. The specimens are in different 3D positions, different orientations, and different developmental and degradation stages with wide intra-class variation. On this very challenging data set, our new algorithm outperforms other classifiers, showing the benefits of using spatial information in the stacking framework with multiple dissimilar feature types.

Responses of Litter-Dwelling Arthropods to four Different Thinning Intensities in Douglas-Fir Forests of the Pacific Northwest, USA

We investigated the response of litter-dwelling arthropods to the effects of four forestry thinning intensities (Control, Light Thin, Light Thin with Gap, and Heavy Thin). With the balance between timber demand and maintaining biodiversity in the forest ecosystem in mind, we examined the effects of thinning on the abundance, richness, and diversity of arthropods as an indicator of how ecological processes affect forest litter-dwelling fauna. Study sites were 40- to 60-year-old stands of typical Douglas-fir plantation in the Willamette National Forest, Oregon, USA. To examine the seasonal response of the litter-dwelling arthropods, litter debris and humus samples were collected in October 2000 (wet late-growing season, Late 2000), June 2001 (wet early-growing season, Early 2001), and August 2001 (dry mid-growing season, Mid 2001) and extracted with Tullgren funnels. The abundance and diversity of litter-dwelling arthropods decreased as thinning intensity increased. The decreases in both abundance and diversity of arthropods with limited mobility within the two heaviest thinnings were correlated with an increased heterogeneity of disturbance to the forest floor (patchy litter and moss cover removal), rather than responses to thinning itself at the scale of the entire stand. The litter-dwelling fauna correlated positively with litter moisture. Under control conditions, the abundance of predators and detritivores increased during the dry summer in August. Non-metric multidimensional scaling results showed distinct clusters for the three growing seasons. The wet early-growing season clustered with the dry mid-growing season, but not with the wet late-growing season. Moisture correlated highly with the ordination axes. This study showed that litter-dwelling arthropods were correlated with stand density and seasonal litter moisture of the forest floor.

AN AQUATIC INSECT IMAGING SYSTEM TO AUTOMATE INSECT CLASSIFICATION

Population counts of aquatic insects are a valuable tool for monitoring the water quality of rivers and streams. However, the handling of samples in the lab for species identification is time consuming and requires specially trained experts. An aquatic insect imaging system was designed as part of a system to automate aquatic insect classification and was tested using several species and size classes of stonefly (Plecoptera). The system uses ethanol to transport specimens via a transparent rectangular tube to a digital camera. A small jet is used to position and reorient the specimens so that sufficient pictures can be taken to classify them with pattern recognition. A mirror system is used to provide a split set of images 90° apart. The system is evaluated with respect to engineering requirements developed during the research, including image quality, specimen handling, and system usability.