Capella Partners with IEEE GRSS for 2026 Data Fusion Contest

After nearly 40 years, an idea inspired by Professor Jon Claerbout’s lectures in Tokyo has finally taken shape as a paper. The work revisits the long-standing free-surface ghost problem in marine seismic data from a different perspective — treating ghosts not as artifacts to remove, but as mirror observations that can be constructively exploited through processing. The result is a processing-driven framework that avoids explicit inverse ghost filtering and offers a pathway for broadband recovery, particularly for legacy marine datasets. The preprint is now available on arXiv: https://lnkd.in/gNycDX4Q It is deeply satisfying to finally bring this long-standing idea to completion. I look forward to discussions and perspectives from colleagues working in marine seismic processing and broadband acquisition.

My First Publication 🗿 🤖 : Application of Transformer Models for Volcano Seismic Signals Classification DOI: 10.1109/ColCACI67437.2025.11230894 https://lnkd.in/dr5x5VCF Publisher: IEEE For me, it was especially meaningful because I used a transformer model for this work and learned a lot about how deep learning really works. It is fascinating to see how you can create a model that learns from data and can help solve real-world problems like volcanic monitoring. This paper is interesting to me because I proposed an automatic volcanic seismic event classification method based on transformer models that works directly on raw signals, avoiding complex preprocessing. Using the MicSigv1 dataset from Cotopaxi volcano, the model achieved around 96% in F1-score and accuracy.

Over the past year, the group composed of researchers from the Digital Technologies and Cities, Infrastructure and Environment Units of the Instituto de Pesquisas Tecnológicas, together with researchers from Petrobras, had three papers accepted and presented at major national and international conferences, all focused on the application of Artificial Intelligence and Machine Learning to the interpretation of seismic data. The studies ranged from systematic reviews supported by generative models to advanced methods for the automatic detection of geological features in 3D seismic volumes, using multiple attributes and segmentation architectures. ✅ Identifying AI Techniques for Detecting Geological Features in Seismic Data: Literature Review Method with GPT-4 (86th EAGE Annual Conference & Exhibition, Toulouse – France), available at: https://lnkd.in/dqFdcQ6W ✅ Fault Detection in 3D Seismic Data with Machine Learning Methods and Multiple Attributes (19th Congress of the Brazilian Geophysical Society, Rio de Janeiro – Brazil), available at: https://lnkd.in/dGqy7fnZ ✅ Detecting geological features in seismic data using Segment Anything Model 2 across multiple datasets (International Meeting for Applied Geosciences & Energy, Houston – USA), available at: https://lnkd.in/d4J5BPkD These contributions reinforce our role in developing innovative solutions for geophysical analysis, integrating data science, automation, and geological interpretation across different contexts. I invite all interested readers to access the full papers through the links above.

Dr. Hang Chen has published a research article in Geophysical Research Letters with his PhD student Chen Xiong as one of the co-authors. The paper outlines a Model-Experiment (ModEx) framework that integrates limited EMI data with machine learning (ML) and hydrologic modeling to improve process representation and guide field investigations. Read the article here: https://lnkd.in/g4eYCbE7

This week, GLOS’ observing team is on the road at two different meetings. Shelby Brunner, Science and Observations Manager, is presenting on the evolution of the Great Lakes observing network at the 2026 Ocean Sciences Meeting in Glasgow, Scotland. Shelby is looking forward to hearing talks on the most cutting edge “large water” science, to help improve capacity in the region and network. Connor B., Observing Data Specialist, is participating in the GLATOS Annual Meeting in Sarnia, Ontario, to share GLOS’ project for augmenting the Great Lakes Acoustic Telemetry Observation System (GLATOS) moorings with additional sensors to expand availability of year-round data. #greatlakesobservingsystem #yeahbuoy

📣 New Seismic Soundoff episode! High performance computing is no longer optional. Madhav Vyas and Elizabeth L’Heureux share insights on computation, algorithms, and the skills shaping the future of geophysics. Tune in: https://lnkd.in/geZsEEKC

Explore. Analyse. Invest. The Virtual Core Library (VCL) by the Council for Geoscience gives you digital access to high-quality drill core imagery and geological data anytime, anywhere. Watch full video on YouTube: https://lnkd.in/dWGQnz_T

Great to see South Africa's Council for Geoscience utilizing TerraCore's powerful data analysis software, IntelliCore. In addition to hyperspectral data, you can add any data (ex: logs, geochem) to IntelliCore so you can compare it downhole or versus imagery. If you are headed to Prospectors & Developers Association of Canada (PDAC) in Toronto next week, stop by Booth 1324 to check it out live! #hyperspectral #MiningTechnology #MiningInnovation #PDAC2026

Very proud of you Clement Ndou (Pr.Sci.Nat) and the entire team on successfully bringing the Virtual Core Library to life. This milestone reflects not only technical excellence and innovation, but also a strong commitment to advancing collaboration and modernising how we access, interpret, and share geoscientific data. Initiatives like this continue to push our industry forward and create meaningful impact across the geoscience community. Proud to see such leadership and vision turning ideas into reality.

Our latest paper is now available #openaccess in Earth-Science Reviews. In this study, we use GNSS data, machine-learning clustering, and 3D strain-rate analysis to explore how glacial isostatic adjustment contributes to intraplate deformation and seismicity in Eastern Canada, and what that means for seismic hazard assessment. The paper is here: https://lnkd.in/eSJeSn6E McMaster University McMaster Research McMaster Faculty of Science Sichuan University Virginia Tech