A5036360490- Species Distribution and Climate Change
- Advanced Image and Video Retrieval Techniques
- Advanced Vision and Imaging
- Wildlife Ecology and Conservation
- Smart Agriculture and AI
- Generative Adversarial Networks and Image Synthesis
- Image Processing Techniques and Applications
- Image Processing and 3D Reconstruction
- Remote Sensing and LiDAR Applications
- Plant Pathogens and Fungal Diseases
- Environmental DNA in Biodiversity Studies
- Music and Audio Processing
- Video Analysis and Summarization
- Video Surveillance and Tracking Methods
- Domain Adaptation and Few-Shot Learning
- Cell Image Analysis Techniques
- Postharvest Quality and Shelf Life Management
- Insect and Arachnid Ecology and Behavior
- Horticultural and Viticultural Research
- Image Enhancement Techniques
- Rangeland and Wildlife Management
- Advanced Image Processing Techniques
Massachusetts Institute of Technology
2024
University of Bonn
2021-2024
Vassar College
2024
Rapid changes of the biosphere observed in recent years are caused by both small and large scale drivers, like shifts temperature, transformations land-use, or energy budget systems. While latter processes easily quantifiable, documentation loss biodiversity community structure is more difficult. Changes organismal abundance diversity barely documented. Censuses species usually fragmentary inferred often spatially, temporally ecologically unsatisfactory simple lists for individual study...
Accurate estimates of salmon escapement - the number fish migrating upstream to spawn are key data for conservation and fishery management. Existing methods counting using high-resolution imaging sonar hardware non-invasive compatible with computer vision processing. Prior work in this area has utilized object detection tracking based automated counting. However, these techniques remain inaccessible many deployment sites due limited compute connectivity field. We propose an alternative...
Abstract Camera traps have become important tools for the monitoring of animal populations. However, study‐specific estimation detection probabilities is key if unbiased abundance estimates unmarked species are to be obtained. Since this process can very time‐consuming, we developed first semi‐automated workflow animals any size and shape estimate population densities. In order obtain observation distances, a deep learning algorithm used create relative depth images that calibrated with...
Camera traps are a proven tool in biology and specifically biodiversity research. However, camera including depth estimation not widely deployed, despite providing valuable context about the scene facilitating automation of previously laborious manual ecological methods. In this study, we propose an automated trap-based approach to detect identify animals using estimation. To individual animals, novel method D-Mask R-CNN for so-called instance segmentation which is deep learning-based...
The development and application of modern technology are an essential basis for the efficient monitoring species in natural habitats to assess change ecosystems, communities populations, order understand important drivers change. For estimating wildlife abundance, camera trapping combination with three-dimensional (3D) measurements is highly valuable. Additionally, 3D information improves accuracy detection using trapping. This study presents a novel approach featuring optimized hardware...
Grape resilience towards Botrytis cinerea (B. cinerea) infections (Botrytis bunch rot) is an important concern of breeders and growers. Beside grape architecture, berry surface characteristics like bloom (epicuticular wax) as well thickness permeability the cuticle represent further promising physical barriers to increase rot. In previous studies, two efficient sensor-based phenotyping methods were developed evaluate both traits fast objectively: (1) light-separated RGB (red-green-blue)...
Camera traps, an invaluable tool for biodiversity monitoring, capture wildlife activities day and night. In low-light conditions, near-infrared (NIR) imaging is commonly employed to images without disturbing animals. However, the reflection properties of NIR light differ from those visible in terms chrominance luminance, creating a notable gap human perception. Thus, objective enrich with colors, thereby bridging this domain gap. Conventional colorization techniques are ineffective due...
Object detectors often perform poorly on data that differs from their training set. Domain adaptive object detection (DAOD) methods have recently demonstrated strong results addressing this challenge. Unfortunately, we identify systemic benchmarking pitfalls call past into question and hamper further progress: (a) Overestimation of performance due to underpowered baselines, (b) Inconsistent implementation practices preventing transparent comparisons methods, (c) Lack generality outdated...
The ongoing biodiversity crisis calls for accurate estimation of animal density and abundance to identify sources decline effectiveness conservation interventions. Camera traps together with methods are often employed this purpose. necessary distances between camera observed animals traditionally derived in a laborious, fully manual or semi-automatic process. Both approaches require reference image material, which is both difficult acquire not available existing datasets. We propose...
The development and application of modern technology is an essential basis for the efficient monitoring species in natural habitats landscapes to trace ecosystems, communities, populations, analyze reasons changes. For estimating animal abundance using methods such as camera trap distance sampling, spatial information terms 3D (three-dimensional) measurements crucial. Additionally, improves accuracy detection trapping. This study presents a novel approach trapping featuring highly optimized...