Automatic Fungi Recognition: Deep Learning Meets Mycology

The article presents an AI-based fungi species recognition system for a citizen-science community. The system’s real-time identification too — FungiVision — with a mobile application front-end, led to increased public interest in fungi, quadrupling the number of citizens collecting data. FungiVision, deployed with a human-in-the-loop, reaches nearly 93% accuracy. Using the collected data, we developed a novel fine-grained classification dataset — Danish Fungi 2020 (DF20) — with several unique characteristics: species-level labels, a small number of errors, and rich observation metadata. The dataset enables the testing of the ability to improve classification using metadata, e.g., time, location, habitat and substrate, facilitates classifier calibration testing and finally allows the study of the impact of the device settings on the classification performance. The continual flow of labelled data supports improvements of the online recognition system. Finally, we present a novel method for the fungi recognition service, based on a Vision Transformer architecture. Trained on DF20 and exploiting available metadata, it achieves a recognition error that is 46.75% lower than the current system. By providing a stream of labeled data in one direction, and an accuracy increase in the other, the collaboration creates a virtuous cycle helping both communities.

DBMF: A Novel Method for Tree Species Fusion Classification Based on Multi-Source Images

Multi-source data remote sensing provides innovative technical support for tree species recognition. Tree species recognition is relatively poor despite noteworthy advancements in image fusion methods because the features from multi-source data for each pixel in the same region cannot be deeply exploited. In the present paper, a novel deep learning approach for hyperspectral imagery is proposed to improve accuracy for the classification of tree species. The proposed method, named the double branch multi-source fusion (DBMF) method, could more deeply determine the relationship between multi-source data and provide more effective information. The DBMF method does this by fusing spectral features extracted from a hyperspectral image (HSI) captured by the HJ-1A satellite and spatial features extracted from a multispectral image (MSI) captured by the Sentinel-2 satellite. The network has two branches in the spatial branch to avoid the risk of information loss, of which, sandglass blocks are embedded into a convolutional neural network (CNN) to extract the corresponding spatial neighborhood features from the MSI. Simultaneously, to make the useful spectral feature transfer more effective in the spectral branch, we employed bidirectional long short-term memory (Bi-LSTM) with a triple attention mechanism to extract the spectral features of each pixel in the HSI with low resolution. The feature information is fused to classify the tree species after the addition of a fusion activation function, which could allow the network to obtain more interactive information. Finally, the fusion strategy allows for the prediction of the full classification map of three study areas. Experimental results on a multi-source dataset show that DBMF has a significant advantage over other state-of-the-art frameworks.

Phylogenetic Relationships, Speciation, and Origin of Armillaria in the Northern Hemisphere: A Lesson Based on rRNA and Elongation Factor 1-Alpha

Armillaria species have a global distribution and play various roles in the natural ecosystems, e.g., pathogens, decomposers, and mycorrhizal associates. However, their taxonomic boundaries, speciation processes, and origin are poorly understood. Here, we used a phylogenetic approach with 358 samplings from Europe, East Asia, and North America to delimit the species boundaries and to discern the evolutionary forces underpinning divergence and evolution. Three species delimitation methods indicated multiple unrecognized phylogenetic species, and biological species recognition did not reflect the natural evolutionary relationships within Armillaria; for instance, biological species of A. mellea and D. tabescens are divergent and cryptic species/lineages exist associated with their geographic distributions in Europe, North America, and East Asia. While the species-rich and divergent Gallica superclade might represent three phylogenetic species (PS I, PS II, and A. nabsnona) that undergo speciation. The PS II contained four lineages with cryptic diversity associated with the geographic distribution. The genus Armillaria likely originated from East Asia around 21.8 Mya in early Miocene when Boreotropical flora (56–33.9 Mya) and the Bering land bridge might have facilitated transcontinental dispersal of Armillaria species. The Gallica superclade arose at 9.1 Mya and the concurrent vicariance events of Bering Strait opening and the uplift of the northern Tibetan plateau might be important factors in driving the lineage divergence.

Morphometric and genetic identification of a newly record pygmy seahorse Hippocampus denise in the Kepulauan Seribu reefs, Indonesia

The study was conducted to describe the seahorse species based on morphological and molecular characters. The pygmy seahorse in Panggang Island in Kepualuan Seribu was discovered in October 2011. The species was allegedly identified as Hippocampus denise (Family: Syngnathidae) described by Lourie and Randall which published in 2003. The high similarity is based on small morphometric, orange-like color and its association with sea fan Annella sp. Their habitat is fairly shallow at a depth between 13-24 meters compared with their sister species observed in Bali, Nusa Tenggara, and Sulawesi. The phylogenetic analysis constructed with several sequence data of Hippocampus spp. from Genbank shows that sample collected from Panggang Island is in the same clade with Hippocampus denise with 100% bootstrap value. BLAST analysis result also showed a high maximum similar identity (>99%) with the species Hippocampus denise. The seahorse specimen described in this study has a common typology of habitat with Hippocampus denise. This study shows that genetic analysis to determine the Hippocampus denise can be carried out to support species recognition, especially for cryptic species such as Hippocampus spp. There are variations in morphometric and habitat depth levels, indicating local adaptation of pygmy seahorses to the Kepulauan Seribu reefs.

PLRSNet: a semantic segmentation network for segmenting plant leaf region under complex background

PurposeExtraction of leaf region from the plant leaf images is a prerequisite process for species recognition, disease detection and classification and so on, which are required for crop management. Several approaches were developed to implement the process of leaf region segmentation from the background. However, most of the methods were applied to the images taken under laboratory setups or plain background, but the application of leaf segmentation methods is vital to be used on real-time cultivation field images that contain complex backgrounds. So far, the efficient method that automatically segments leaf region from the complex background exclusively for black gram plant leaf images has not been developed.Design/methodology/approachExtracting leaf regions from the complex background is cumbersome, and the proposed PLRSNet (Plant Leaf Region Segmentation Net) is one of the solutions to this problem. In this paper, a customized deep network is designed and applied to extract leaf regions from the images taken from cultivation fields.FindingsThe proposed PLRSNet compared with the state-of-the-art methods and the experimental results evident that proposed PLRSNet yields 96.9% of Similarity Index/Dice, 94.2% of Jaccard/IoU, 98.55% of Correct Detection Ratio, Total Segmentation Error of 0.059 and Average Surface Distance of 3.037, representing a significant improvement over existing methods particularly taking into account of cultivation field images.Originality/valueIn this work, a customized deep learning network is designed for segmenting plant leaf region under complex background and named it as a PLRSNet.

Neural Basis of Acoustic Species Recognition in a Cryptic Species Complex

Sexual traits that promote species recognition are important drivers of reproductive isolation, especially among closely related species. Identifying neural processes that shape species differences in recognition is crucial for understanding the causal mechanisms of reproductive isolation. Temporal patterns are salient features of sexual signals widely used in species recognition by several taxa, including anurans. Recent advances in our understanding of temporal processing by the anuran auditory system provide an opportunity to investigate the neural basis of species-specific recognition. The anuran inferior colliculus (IC) consists of neurons that are selective for temporal features of calls. Of potential relevance are auditory neurons known as interval-counting neurons (ICNs) that are often selective for the pulse rate of conspecific advertisement calls. Here, we tested the hypothesis that ICNs mediate acoustic species recognition by exploiting the known differences in temporal selectivity in two cryptic species of gray treefrog (Hyla chrysoscelis and Hyla versicolor). We examined the extent to which the threshold number of pulses required to elicit behavioral responses from females and neural responses from ICNs was similar within each species but potentially different between the two species. In support of our hypothesis, we found that a species difference in behavioral pulse number thresholds closely matched the species difference in neural pulse number thresholds. However, this relationship held only for ICNs that exhibited band-pass tuning for conspecific pulse rates. Together, these findings suggest that differences in temporal processing of a subset of ICNs provide a mechanistic explanation for reproductive isolation between two cryptic treefrog species.

Additions to Philippine slender skinks of the Brachymeles bonitae complex (Reptilia: Squamata: Scincidae) IV: Resurrection and redescription of Brachymeles burksi

The diversity of Philippine amphibians and reptiles has increased over the last few decades, in part due to re-evaluation of species formerly believed to be widespread. Many of these investigations of widespread species have uncovered multiple closely related cryptic lineages comprising species complexes, each restricted to individual Pleistocene Aggregate Island Complexes (PAICs). One group in particular for which widespread cryptic diversity has been common is the clade of Philippine skinks of the genus Brachymeles. Recent phylogenetic studies of the formerly recognized widespread species Brachymeles bonitae have indicated that this species is actually a complex distributed across several major PAICs and smaller island groups in the central and northern Philippines, with numerous species that exhibit an array of digit loss and limb reduction patterns. Despite the recent revisions to the B. bonitae species complex, studies suggest that unique cryptic lineages still exist within this group. In this paper, we resurrect the species Brachymeles burksi Taylor 1917, for a lineage of non-pentadactyl, semi-fossorial skink from Mindoro and Marinduque islands. First described in 1917, B. burksi was synonymized with B. bonitae in 1956, and has rarely been reconsidered since. Evaluation of genetic and morphological data (qualitative traits, meristic counts, and mensural measurements), and comparison of recently-obtained specimens to Taylor’s original description support this species’ recognition, as does its insular distribution on isolated islands in the central portions of the archipelago. Morphologically, B. burksi is differentiated from other members of the genus based on a suite of unique phenotypic characteristics, including a small body size, digitless limbs, a high number of presacral vertebrae, the absence of auricular openings, and discrete (non-overlapping) meristic scale counts. The recognition of this central Philippine species further increases the diversity of non-pentadactyl members of the B. bonitae complex, and reinforces the biogeographic uniqueness of the Mindoro faunal region. KEYWORDS: biodiversity, endemism, faunal region, fossoriality, limb reduction