We introduce EyeTheia, a lightweight and open deep learning pipeline for webcam-based gaze estimation, designed for browser-based experimental platforms and real-world cognitive and clinical research. EyeTheia enables real-time gaze tracking using only a standard laptop webcam, combining MediaPipe-based landmark extraction with a convolutional neural network inspired by iTracker and optional user-specific fine-tuning. We investigate two complementary strategies: adapting a model pretrained on mobile data and training the same architecture from scratch on a desktop-oriented dataset. Validation results on MPIIFaceGaze show comparable performance between both approaches prior to calibration, while lightweight user-specific fine-tuning consistently reduces gaze prediction error. We further evaluate EyeTheia in a realistic Dot-Probe task and compare it to the commercial webcam-based tracker SeeSo SDK. Results indicate strong agreement in left–right gaze allocation during stimulus presentation, despite higher temporal variability. Overall, EyeTheia provides a transparent and extensible solution for low-cost gaze tracking, suitable for scalable and reproducible experimental and clinical studies. The code, trained models, and experimental materials are publicly available ( EyeTheia repository and Calypso experimental platform ).

Automatic surgical gesture recognition has the potential to revolutionize the field of surgery by enhancing patient care, surgical training, and our understanding of surgical skills. By integrating kinematic data, which precisely captures hand movements, with video data for contextual understanding, multimodal machine learning can greatly enhance the accuracy of surgical gesture recognition systems by capturing complementary knowledge. Recent research has highlighted the capabilities of Transformer-based models for temporal action segmentation. A key component of these models is the iterative refinement module, which enhances predictions using contextual data. In this study, we propose MGRFormer, a novel multimodal framework that leverages the interaction between kinematics and visual data at the refinement stage for the task of surgical gesture recognition. We evaluated our MGRFormer on the VTS dataset, and the results demonstrated that our approach outperformed unimodal and multimodal state-of-the-art methods by a large margin.

This paper aims to understand the practical features of hierarchies of morphological segmentations, namely the quasi-flat zones hierarchy and watershed hierarchies, and to evaluate their potential in the context of natural image analysis. We propose a novel evaluation framework for the hierarchies of partitions designed to capture various aspects of those representations: precision of their regions and contours, possibility to extract high quality horizontal cuts and optimal non-horizontal cuts for image segmentation, and the ease of finding a set of regions representing a semantic object. This framework is used to assess and to optimize hierarchies with respect to the possible pre- and post-processing steps. We show that, used in conjunction with a state-of-the-art contour detector, watershed hierarchies are competitive with the complex state-of-the-art methods for hierarchy construction. In particular, the proposed framework allows us to identify a watershed hierarchy based on a novel extinction value, the number of parent nodes that outperforms the other hierarchies of morphological segmentations. This coupled with the fact that watershed hierarchies satisfy clear global optimality properties and can be efficiently computed on large data, make them valuable candidates for various computer vision tasks.