Rafael M. O. Cruz

Eduardo V. L. Barboza, Jean Paul Barddal, Robert Sabourin et al.

This work presents Causal Drift Generator (CaDrift), a time-dependent synthetic data generator framework based on Structural Causal Models (SCMs). The framework produces a virtually infinite combination of data streams with controlled shift events and time-dependent data, making it a tool to evaluate methods under evolving data. CaDrift synthesizes various distributional and covariate shifts by drifting mapping functions of the SCM, which change underlying cause-and-effect relationships between features and the target. In addition, CaDrift models occasional perturbations by leveraging interventions in causal modeling. Experimental results show that, after distributional shift events, the accuracy of classifiers tends to drop, followed by a gradual retrieval, confirming the generator's effectiveness in simulating shifts. The framework has been made available on GitHub.

Kecia G. de Moura, Robert Sabourin, Rafael M. O. Cruz

Offline handwritten signature verification aims to distinguish genuine from forged signatures using static images. Since real forgeries are rarely available, negative samples are usually randomly drawn from genuine signatures of other users to create training data. However, this random selection often lacks diversity, increases redundancy, and escalates computational cost, leading to inefficient training. We propose a data-driven strategy to generate diverse, informative negative samples using prototypical signatures, which are compact, non-identifiable summaries of genuine signature features. Based on the experiments results, we conclude that (i) prototypical signatures yield more informative negative samples, improving the detection of skilled forgeries; (ii) the proposed approach is backbone-agnostic, showing robustness across architectures; and (iii) when combined with a primal-form linear SVM, it serves as an alternative to RBF-based models while significantly improving scalability and computational efficiency. Implementation of the method is available at https://github.com/kdmoura/proto_hsv.

Arthur Josi, Emeline Got, Abdallah Dib et al.

Speech-driven 3D facial animation methods face significant challenges in simultaneously achieving high-fidelity motion and precise artistic control at production quality. Existing controllable models typically learn global style control by relying on large-scale, low-quality \emph{in-the-wild} datasets that compromise overall animation realism. Furthermore, these frameworks often lack the fine-grained temporal precision required for demanding tasks such as dialogue localization (e.g., dubbing), where matching specific facial expressions is as critical as lip synchronization. We present KM-Speaker (Keypoint-Matching Speaker), a novel keypoint-conditioned flow-based generative framework that provides both global style guidance and frame-level temporal control from reference performances. We propose a disentanglement strategy that separates audio-driven lip motion from keypoint-driven upper-face dynamics, together with a global style context preservation mechanism to ensure coherent full-face expressiveness. KM-Speaker advances example-based 3D facial animation by achieving high-fidelity motion and flexible controllability in a data-constrained setting, consistently outperforming state-of-the-art methods in lip-sync accuracy, style adherence, and expressive temporal control.

Eduardo V. L. Barboza, Robert Sabourin, Rafael M. O. Cruz

The k-Nearest Neighbors (kNN) algorithm has long been widely used in Machine Learning (ML) applications. However, the main concern when using it is the computational cost required for neighborhood search, which can make it unfeasible for large-scale applications. Optimization algorithms, such as the K-d tree, become an option in such scenarios. Under data streams, it can be challenging to maintain the properties of the K-d tree, as it requires inserting and deleting nodes on the fly. These operations can make maintaining the tree's balance and invariants difficult. Additionally, traditional K-d trees were initially designed for Minkowski-based distance functions. In this work, we describe an Online K-d tree and its adaptation to the Canberra distance that supports dynamic updates over data streams while preserving the structural invariants required for efficient traversal. Experimental analysis demonstrates that the Online K-d tree algorithm achieves faster processing time under data streams, and that adapting to the Canberra distance enabled effective subtree pruning, as evidenced by a minor loss in average accuracy and a substantial gain in instances processed per second. Our implementation can be found in our GitHub repository