Project ATHENA

Project Overview

The ATHENA project develops Artificial Intelligence (AI) technologies to improve minimally invasive oncological surgery (laparoscopy). This directly addresses the challenge of "Health, Demographic Change, and Well-being."

The main challenge in these surgeries is the difficulty of locating tumors hidden in deformable organs (such as the liver, kidney, or uterus), as the surgeon loses tactile and 3D perception. ATHENA's solution is based on Augmented Reality (AR), overlaying a 3D map of the organ and tumors onto the surgeon's video feed.

Key Innovation: We have developed novel Deep Learning algorithms to solve the complex problem of aligning a 3D preoperative map to a moving, deforming organ in real-time, without the need for complex synthetic data or external sensors.

Shape-from-Template (SfT) and Non-Rigid Structure-from-Motion (NRSfM) diagrams

Figure: Concept of SfT and NRSfM applied to surgery.

Scientific Objectives

General Representations

Study representations of templates, objects, and deformations capable of interacting realistically with Deep Neural Networks (DNN), focusing on thin-shell and volumetric objects.

General DNN Architecture

Define a general DNN architecture to solve Shape-from-Template (SfT) and Non-Rigid Structure-from-Motion (NRSfM) end-to-end.

Self-Supervised Learning

Develop methods to train DNNs using unlabeled images of deforming objects, utilizing optical flow and photometric error minimization.

Standardization & Datasets

Creation of synthetic and real (RGBD) datasets to standardize the evaluation of 3D deformable reconstruction in the scientific community.

AR Laparoscopy

Propose viable solutions for Augmented Reality guided laparoscopy for soft organs (specifically the liver) using the developed DNN methods.

Results & Publications

Key Journal Publications

[R1] Deep Shape-from-Template: Single-image quasi-isometric deformable registration and reconstruction.
Fuentes-Jimenez, D., Pizarro, D., et al.
Image and Vision Computing, 127, 104531 (2022). DOI
[R2] Label Augmentation to Improve Generalization of Deep Learning Semantic Segmentation of Laparoscopic Images.
Monasterio-Exposito, L., Pizarro, D., et al.
IEEE Access, 10, 37345-37359 (2022). DOI
[R3] Neural patient-specific 3d–2d registration in laparoscopic liver resection.
Mhiri, I., Pizarro, D., & Bartoli, A.
International Journal of Computer Assisted Radiology and Surgery, 20(1), 57-64 (2025).
[R4] Weakly-Supervised Deep Shape-From-Template.
S. Luengo-Sanchez, D. Fuentes-Jimenez, et al.
IEEE Access, vol. 13, pp. 22868-22892 (2025).
[R5] MIS-NeRF: neural radiance fields in minimally-invasive surgery.
Khojasteh, S. B., et al.
International Journal of Computer Assisted Radiology and Surgery (2025).
[R6] Detection of Anomalies in Daily Activities Using Data from Smart Meters.
A. Hernández, R. Nieto, et al.
Sensors, vol. 24(515), pp. 1-17 (2024).

Books & Book Chapters

[C1] Automatic 3D/2D Deformable Registration in Minimally Invasive Liver Resection using a Mesh Recovery Network.
Labrunie, M., Pizarro, D., Tilmant, C., & Bartoli, A.
Medical Imaging with Deep Learning (pp. 1104-1123). PMLR (Jan 2024).
[C2] Generic Liver Modelling with Application to Mini-invasive Surgery Guidance.
Labrunie, M., Pizarro, D., Tilmant, C., & Bartoli, A.
International Conference on Medical Imaging and Computer-Aided Diagnosis (pp. 219-229). Springer Nature Singapore (Nov 2024).

Conference Proceedings & Technical Publications

[C3] Estimating Energy Consumption in Households for Non-Intrusive Elderly Monitoring.
A. Hernández, L. de Diego, D. Pizarro, et al.
IEEE MetroLivEnv 2023 Proceedings, pp. 1-5 (Milano, Italy, 2023).
[C4] Appliance Identification in NILM Applications by means of a Convolutional Auto-Encoder.
L. de Diego, A. Hernández, D. Pizarro, R. Nieto.
IEEE MetroLivEnv 2023 Proceedings, pp. 1-6 (Milano, Italy, 2023).
[C5] Implementing a CNN in FPGA Programmable Logic for NILM Application.
M. Tapiador, L. de Diego-Otón, A. Hernández, R. Nieto.
Proceedings of 38th Conference on Design of Circuits and Integrated Systems (DCIS 2023).
[C6] Comparative Analysis of Neural Network Implementations for NILM Applications.
J. Martín, L. de Diego-Otón, M. Tapiador, A. Hernández, R. Nieto.
Proceedings of 38th Conference on Design of Circuits and Integrated Systems (DCIS 2023).
[C7] SoC Architecture for High-Frequency Acquisition of Household Electric Signals.
V. Navarro, L. de Diego, M. Tapiador, R. Nieto, J. Ureña, A. Hernández.
Proceedings of 2024 IEEE International Instrumentation and Measurement Technology Conference (I2MTC 2024).
[C8] Preliminary Unknown Appliance Detection using Convolutional Variational Auto-Encoders for AAL.
L. de Diego-Otón, D. Fuentes, D. Pizarro, A. Hernández, et al.
Proceedings of IEEE International Conference on Omni-layer Intelligent Systems (COINS 2024).
[C9] An Open-Source VLBI Digital Backend for Low-Cost FPGA-based SoCs.
M. Cubero, A. Hernández, J. González.
Proceedings of 39th Conference on Design of Circuits and Integrated Systems (DCIS 2024).

PhD Theses Defended

David Fuentes Jiménez - "Reconstrucción de objetos deformables a partir de imágenes mediante técnicas de deep learning." (2021)
Leticia Monasterio Expósito - "Segmentación Semántica Interactiva en Imágenes de Laparoscopia mediante Redes Neuronales Profundas." (2023)
Mathieu Labrunie - "Contributions to the automatic registration of a 3D preoperative model with a 2D image in mini-invasive surgery of the liver." (2024)

Project Overview

Scientific Objectives

General Representations

General DNN Architecture

Self-Supervised Learning

Standardization & Datasets

AR Laparoscopy

Results & Publications

Key Journal Publications

Books & Book Chapters

Conference Proceedings & Technical Publications

PhD Theses Defended

The Team

Daniel Pizarro Pérez

Research Team

Álvaro Hernández Alonso

Manuel Sánchez Chapado

José Luis Martín Sánchez

Cristina Losada Gutiérrez

Work Team & Collaborators

Adrien Bartoli

Leticia Monasterio

David Fuentes Jiménez

Sara Luengo Sánchez

Samad Barri Khojasteh

Partners & Funding

Universidad de Alcalá

EnCoV

SurgAR