S. Marconi

Individual tree crown maps for the National Ecological Observatory Network

The paper uses computer vision and deep neural networks to predict individual tree species, crown characteristics, and locations across 24 sites in the National Ecological Observatory Network.

Capturing long-tailed individual tree diversity using an airborne multi-temporal hierarchical model

The paper develops a multi-temporal hierarchical model to improve rare tree species classification in airborne imagery, achieving 75% accuracy for 14 species and generating species maps for over 670,000 trees.

Quantitative Assessment of 3D Printed Blood Vessels Produced with J750™ Digital Anatomy™ for Suture Simulation

This paper evaluates 3D printed blood vessels using J750™ Digital Anatomy™ to simulate suture performance, identifying optimal material combinations for anastomosis training.

Continental-scale hyperspectral tree species classification in the United States National Ecological Observatory Network

The paper develops a continental-scale tree species classifier using NEON data, showing better accuracy than site-specific models and highlighting challenges in species classification across diverse ecosystems.

Data science competition for cross-site delineation and classification of individual trees from airborne remote sensing data

This paper runs a data science competition to improve methods for delineating and classifying individual trees from remote sensing data across multiple sites.

Disentangling the roles of inter and intraspecific variation on leaf trait distributions across the eastern United States

The paper develops a joint model to analyze how inter- and intraspecific variation influence leaf trait distributions across the eastern US, finding that intraspecific variation significantly contributes to trait variation.

Estimating individual level plant traits at scale

This paper develops a method to estimate individual tree traits at scale using remote sensing data, achieving moderate accuracy for several traits and demonstrating improved performance over pixel-level predictions.

Individual tree-crown detection in RGB imagery using semi-supervised deep learning neural networks

This paper develops a semi-supervised deep learning method to detect tree crowns in RGB imagery using unsupervised LIDAR data and a small number of hand-annotated images, achieving accurate predictions in natural landscapes.

A benchmark dataset for individual tree crown delineation in co-registered airborne RGB, LiDAR and hyperspectral imagery from the National Ecological Observation Network

The paper introduces a benchmark dataset for evaluating individual tree crown delineation algorithms using co-registered remote sensing data and provides an R package for standardized evaluation.

NEON Crowns: a remote sensing derived dataset of 100 million individual tree crowns

The paper introduces a dataset of 100 million individual tree crowns derived from remote sensing and deep learning, enabling large-scale forest analysis.

Rethinking global carbon storage potential of trees. A comment on Bastin et al. 2019

This paper re-evaluates global carbon storage potential of restored forests, finding it significantly lower than a previous estimate due to flawed assumptions in the original study.

Geographic Generalization in Airborne RGB Deep Learning Tree Detection

This paper evaluates the generalization of RGB deep learning tree detection models across different forest types using data from the National Ecological Observatory Network, finding that a universal model performs as well as site-specific models.