Biology subjects

Vogg, R.

Publications and source records attributed to Vogg, R..

3 recordsLinked to original sources

bioRxiv2025

Foraging competence and scrounging tolerance enhance social relationships in a socially tolerant wild primate

Social interactions are crucial for individual health and ultimately fitness, making the choice of social partners particularly important. Previous research has shown that individuals who succeed in foraging tasks often receive increased affiliation from group members. Similarly, in a social learning context, individuals who possess valuable information become more attractive social partners. Thus, an individuals role in a foraging context-specifically, whether it is a successful producer-can influence its social relationships. Therefore, we examined the interplay between social learning, producing and scrounging behavior, and social relationships in four groups of wild redfronted lemurs (Eulemur rufifrons). We conducted an open diffusion experiment with food boxes that required animals to learn one of two techniques to open them. 27 out of 29 individuals participated in the experiment, 24 interacted with the boxes and 16 learned to open them. Initial success was better predicted by use of individual than social information, i.e., manipulating the food boxes vs. time spent watching successful individuals or scrounging. Older males were less successful than females. Scrounging occurred in about 26% of events, with on average 1.3 individuals scrounging. The technique used, age and sex of the producer did not predict scrounging frequency. Learners and males scrounged more often than non-learners and females. Among learners, less successful individuals scrounged more often and this effect was more pronounced in males. More successful individuals and those that were scrounged more often received more affiliative behavior. Thus, cognitive skills and scrounging tolerance may strengthen social relationships in this primate species.

ecology↗

bioRxiv2024

PriMAT: A robust multi-animal tracking model for primates in the wild

O_LIDetection and tracking of animals is an important first step for automated behavioral studies using videos. Animal tracking is currently done mostly using deep learning frameworks based on keypoints, which show remarkable results in lab settings with fixed cameras, backgrounds, and lighting. However, multi-animal tracking in the wild presents several challenges such as high variability in background and lighting conditions, complex motion, and occlusion. C_LIO_LIWe propose a multi-animal tracking model, PriMAT, for nonhuman primates in the wild. The model learns to detect and track primates and other objects of interest from labeled videos or single images using bounding boxes instead of keypoints. Using bounding boxes significantly facilitates data annotation and robustness. Our one-stage model is conceptually simple but highly flexible, and we add a classification branch that allows us to train individual identification. C_LIO_LITo evaluate the performance of our model, we applied it in two case studies with Assamese macaques (Macaca assamensis) and redfronted lemurs (Eulemur rufifrons) in the wild. We show that with only a few hundred frames labeled with bounding boxes, we can achieve robust tracking results. Combining these results with the classification branch for the lemur videos, our model shows an accuracy of 84% in predicting lemur identities. C_LIO_LIOur approach presents a promising solution for accurately tracking and identifying animals in the wild, offering researchers a tool to study animal behavior in their natural habitats. Our code, models, training images, and evaluation video sequences are publicly available1, facilitating their use for animal behavior analyses and future research in this field. C_LI

animal behavior and cognition↗

bioRxiv2022

Using UAV-Imagery for Leaf Segmentation in Diseased Plants via Mask-Based Data Augmentation and Extension of Leaf-based Phenotyping Parameters

AO_SCPLOWBSTRACTC_SCPLOWIn crop production plant diseases cause significant yield losses. Therefore, the detection and scoring of disease occurrence is of high importance. The quantification of plant diseases requires the identification of leaves as individual scoring units. Diseased leaves are very dynamic and complex biological object which constantly change in form and color after interaction with plant pathogens. To address the task of identifying and segmenting individual leaves in agricultural fields, this work uses unmanned aerial vehicle (UAV), multispectral imagery of sugar beet fields and deep instance segmentation networks (Mask R-CNN). Based on standard and copy-paste image augmentation techniques, we tested and compare five strategies for achieving robustness of the network while keeping the number of labeled images within reasonable bounds. Additionally, we quantified the influence of environmental conditions on the network performance. Metrics of performance show that multispectral UAV images recorded under sunny conditions lead to a drop of up to 7% of average precision (AP) in comparison with images under cloudy, diffuse illumination conditions. The lowest performance in leaf detection was found on images with severe disease damage and sunny weather conditions. Subsequently, we used Mask R-CNN models in an image-processing pipeline for the calculation of leaf-based parameters such as leaf area, leaf slope, disease incidence, disease severity, number of clusters, and mean cluster area. To describe epidemiological development, we applied this pipeline in time-series in an experimental trial with five varieties and two fungicide strategies. Disease severity of the model with the highest AP results shows the highest correlation with the same parameter assessed by experts. Time-series development of disease severity and disease incidence demonstrates the advantages of multispectral UAV-imagery for contrasting varieties for resistance, and the limits for disease control measurements. With this work we highlight key components to consider for automatic leaf segmentation of diseased plants using UAV imagery, such as illumination and disease condition. Moreover, we offer a tool for delivering leaf-based parameters relevant to optimize crop production thought automated disease quantification imaging tools.

plant biology↗