Oscar Stanchi
Oscar Stanchi
Computer Engineer and PhD Candidate
I'm currently pursuing a PhD in Computer Science at the Universidad Nacional de La Plata (UNLP) with a scholarship from CONICET. My research focuses on the interpretability of deep neural networks in computer vision, and I primarily work with PyTorch and Captum.
I teach Software Engineering I and II at the Facultad de Informática (UNLP). I also cover topics in Deep Learning, emphasizing its ethical and social implications in the Master's program in Data Mining & Knowledge Discovery, which is offered at the Universidad de Buenos Aires (UBA).
LIDIINFOUNLPCONICETCONICET
Latest Publications
SignAttention: On the Interpretability of Transformer Models for Sign Language Translation

SignAttention: On the Interpretability of Transformer Models for Sign Language Translation
Conference paper

This paper studies a Transformer-based model for translating Greek Sign Language to glosses and text. It finds that the model focuses on clusters of frames, with alignment weakening as glosses increase, and shifts from video frames to predicted tokens during translation. This work enhances understanding of SLT models and promotes more transparent translation systems.

VISTA: Vision Improvement via Split and Reconstruct Deep Neural Network for Fundus Image Quality Assessment

VISTA: Vision Improvement via Split and Reconstruct Deep Neural Network for Fundus Image Quality Assessment
Journal paper

This paper introduces a deep learning model for assessing fundus image quality, essential for diagnosing eye conditions like cataracts and diabetic retinopathy. The model preserves high resolution and includes an autoencoder for reconstruction and classification. Results on the EyeQ dataset show 90.66% accuracy, 88.43% precision, 89.05% recall, and an F1-score of 88.68%.

CB-RISE: Improving the RISE Interpretability Method Through Convergence Detection and Blurred Perturbations

CB-RISE: Improving the RISE Interpretability Method Through Convergence Detection and Blurred Perturbations
Conference paper

This paper enhances the RISE algorithm for image interpretation by introducing C-RISE, which reduces computation time, and CB-RISE, which uses blurred masks instead of black patches. These improvements lead to better heatmap quality and a speedup of about three times.

© 2024 Oscar Stanchi