I'm a Senior Machine Learning Engineer at Pento and a PhD in Computer Vision with over 5 years of experience developing deep learning solutions for image segmentation and analysis.

I completed my PhD thesis on automatic tree-ring detection and analysis, developing novel deep learning architectures and open-source tools now used by the dendrochronology research community worldwide. This work was recognized with First Prize in the Electrical Engineering category by the National Academy of Engineering of Uruguay (ANIU).

• May 2026 TRAS accepted — Forestry, An International Journal of Forest Research
• Dec 2025 First Prize — Doctoral Thesis in Electrical Engineering (ANIU)
• Dec 2025 UruDendro4 presented at ICPRS-25, Viña del Mar, Chile
• Sep 2025 Joined Pento as Senior ML Engineer
• Sep 2025 DeepCS-TRD presented at ICIAP, Rome, Italy
• Aug 2025 PhD completed — Universidad de la República
• 2025 UruDendro dataset family (UruDendro – UruDendro4) published on Zenodo

My research focuses on creating efficient, explainable, and impactful AI systems for challenging domains including dendrochronology, medical imaging, and scientific image analysis. I specialize in deep learning architectures for image segmentation, boundary detection, and real-world image analysis workflows.

All projects fall within the same research line: automatic tree ring detection in wood cross-section images, covering three sub-lines — graphical user interface, computer vision algorithms, and dataset generation.

Graphical User Interface

Computer Vision Algorithms

Dataset Generation

Annotated datasets are mandatory for building tree-ring detection algorithms, are crucial for training and evaluating their performance, and are essential for obtaining accurate results. The datasets are specific for each species under consideration. These datasets were produced in collaboration with multidisciplinary teams, where I contributed as an image processing expert and software developer, working alongside domain specialists with the actual dendrochronological knowledge.

• UruDendro — 64 cross-section images of Pinus taeda L. with annotated annual rings
• UruDendro2 — 53 cross-section images of Pinus taeda L. with earlywood–latewood annotations
• UruDendro3a — 9 cross-section images of Gleditsia triacanthos L.
• UruDendro4 — 102 benchmark images of Pinus taeda L. for ML model development
• Salix glauca — 50 cross-section images from Arctic shrub samples (Greenland)

For a complete list, visit my Google Scholar profile

Recent Journal & Conference Papers

1. Marichal, H., Passarella, D., Randall, G. (2026). TRAS: An Interactive Software for Tracing Tree Ring Cross Sections. Forestry, An International Journal of Forest Research. Accepted.

   Project Page Preprint

2. Marichal, H., Blanco, J., Passarella, D., Randall, G. (2025). UruDendro4: A Benchmark Dataset for Automatic Tree-Ring Detection in Cross-Section Images of Pinus taeda L. IEEE 15th International Conference on Pattern Recognition Systems (ICPRS-25).

   Article Dataset

3. Marichal, H., Casaravilla, V., Power, C., Mello, K., Profumo, L., Mazarino, J., Lucas, C., Passarella, D., Randall, G. (2025). DeepCS-TRD, a Deep Learning-based Cross-Section Tree Ring Detector. International Conference on Image Analysis and Processing (ICIAP).

   Project Page

4. Marichal, H., Passarella, D., Lucas, C., Profumo, L., Casaravilla, V., Rocha Galli, M. N., Ambite, S., Randall, G. (2025). UruDendro, a public dataset of 64 cross-section images and manual annual ring delineations of Pinus taeda L. Annals of Forest Science.

   Article

5. Marichal, H., Passarella, D., Randall, G. (2025). CS-TRD: a Cross-Section Tree Ring Detection Method. Image Processing On Line (IPOL).

   Article

6. Marichal, H., Passarella, D., Randall, G. (2024). Automatic Wood Pith Detector: Local Orientation Estimation and Robust Accumulation. International Conference on Pattern Recognition (ICPR).

   Code

Under Review & Submitted

1. Marichal, H., Power, C., Treier, U. A., Resente, G., Normand, S., Randall, G. (2024). Assessing automatic ring detection on microscopy images of Salix glauca. Under Review.

Thesis

1. Marichal, H. (2025). Application of Image Processing and Artificial Intelligence Techniques for the Automatic Dendrometry of Native and Commercial Wood Species. Doctoral Thesis in Electrical Engineering, Universidad de la República. (Advisors: Gregory Randall and Diego Passarella).

   Read Thesis