Leonardo Cardia da Cruz
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Leonardo Cardia da Cruz has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Alberto Raposo
Tecgraf Institute and Department of Informatics, Pontifical Catholic University of Rio de Janeiro, Rio de Janeiro, Brazil
3 shared publications - Cesar A. Sierra-Franco
Tecgraf Institute and Department of Informatics, Pontifical Catholic University of Rio de Janeiro, Rio de Janeiro, Brazil
3 shared publications - Greis Francy Mireya Silva-Calpa
Pontifical Catholic University of Rio de Janeiro
3 shared publications - Jan Hurtado
Tecgraf Institute and Department of Informatics, Pontifical Catholic University of Rio de Janeiro, Rio de Janeiro, Brazil
2 shared publications - Santiago V. Silva
Tecgraf Institute and Department of Informatics, Pontifical Catholic University of Rio de Janeiro, Rio de Janeiro, Brazil
2 shared publications - Victor de A. Thomaz
Tecgraf Institute and Department of Informatics, Pontifical Catholic University of Rio de Janeiro, Rio de Janeiro, Brazil
2 shared publications
Latest Publications
- Open Access
- 2 March 2026
- Published by Editoriales Iberoamericanos in Revista DCS
- Vol. 23 (88) , e4749
Modelagem conceitual de um sistema inteligente para subsidiar a elaboração de planejamento de aulas e avaliações no ensino superior
A expansão da Educação a Distância e a consolidação da Transformação Digital no ensino superior têm intensificado a necessidade de soluções capazes de integrar tecnologias inteligentes a processos pedagógicos complexos, promovendo maior eficiência, rastreabilidade e consistência metodológica. Nesse cenário, este estudo apresenta um modelo conceitual de sistema inteligente multiagente destinado a automatizar a elaboração do Plano de Aulas e Avaliações (PAA), documento estruturante da gestão pedagógica institucional. A proposta articula fundamentos pedagógicos consolidados - como a Taxonomia de Bloom, sua revisão por Anderson e Krathwohl, e os critérios SMART - a técnicas de Inteligência Artificial, incluindo grandes modelos de linguagem e sistemas multiagentes, organizados em um fluxo human-in-the-loop que preserva a curadoria docente. De natureza qualitativa e caráter documental, a pesquisa organiza uma arquitetura composta por cinco agentes especializados, responsáveis pelo gerenciamento do fluxo informacional, interpretação de diretrizes curriculares, construção das unidades de aprendizagem, validação semântica e pedagógica e geração automatizada do PAA. Os resultados simulados demonstram a coerência do modelo e sua capacidade de estruturar o processo de planejamento em ciclos operacionais iterativos, combinando automação e supervisão humana. A proposta configura uma base conceitual para o desenvolvimento de protótipos futuros, contribuindo para a padronização de documentos instrucionais, para a gestão educacional baseada em dados e para o avanço das discussões sobre o uso da Inteligência Artificial no design instrucional em ambientes digitais.
- 2 March 2026
- Published by Editoriales Iberoamericanos in Revista DCS
- Vol. 23 (88) , e4749
Open AccessModelagem conceitual de um sistema inteligente para subsidiar a elaboração de planejamento de aulas e avaliações no ensino superior
A expansão da Educação a Distância e a consolidação da Transformação Digital no ensino superior têm intensificado a necessidade de soluções capazes de integrar tecnologias inteligentes a processos pedagógicos complexos, promovendo maior eficiência, rastreabilidade e consistência metodológica. Nesse cenário, este estudo apresenta um modelo conceitual de sistema inteligente multiagente destinado a automatizar a elaboração do Plano de Aulas e Avaliações (PAA), documento estruturante da gestão pedagógica institucional. A proposta articula fundamentos pedagógicos consolidados - como a Taxonomia de Bloom, sua revisão por Anderson e Krathwohl, e os critérios SMART - a técnicas de Inteligência Artificial, incluindo grandes modelos de linguagem e sistemas multiagentes, organizados em um fluxo human-in-the-loop que preserva a curadoria docente. De natureza qualitativa e caráter documental, a pesquisa organiza uma arquitetura composta por cinco agentes especializados, responsáveis pelo gerenciamento do fluxo informacional, interpretação de diretrizes curriculares, construção das unidades de aprendizagem, validação semântica e pedagógica e geração automatizada do PAA. Os resultados simulados demonstram a coerência do modelo e sua capacidade de estruturar o processo de planejamento em ciclos operacionais iterativos, combinando automação e supervisão humana. A proposta configura uma base conceitual para o desenvolvimento de protótipos futuros, contribuindo para a padronização de documentos instrucionais, para a gestão educacional baseada em dados e para o avanço das discussões sobre o uso da Inteligência Artificial no design instrucional em ambientes digitais.
- 5 January 2026
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 39 (5) , 4145-4156
Enabling Autonomous Data Annotation in Mammography Image: A Human-in-the-Loop Reinforcement Learning Approach
The vast majority of work in computer vision focuses on developing and applying new machine learning models, which often rely on large amounts of labeled training data. However, annotation is costly and time-consuming. This paper presents an approach based on Deep Reinforcement Learning (DRL) to automatically generate new annotations and reduce human effort in the preparation of training data for supervised learning models in object detection. Our methodology introduces a virtual agent trained with human guidance, inspired by constructivist teaching methods, where interaction with a human teacher supports the agent's learning process. The proposed approach, named "Try a Little More" (TLM), employs active learning to identify uncertain cases and request human intervention during training, progressively enhancing the agent's ability to annotate autonomously. We evaluated our methodology on a mammography dataset, where the agent created bounding box annotations later used in a state-of-the-art supervised object detection algorithm. The results demonstrate that the proposed approach improves the quality and quantity of training data, surpassing existing reinforcement learning and human-computer interaction methods. Quantitatively, TLM generated 414 new annotations with an IoU of 0.86 and F1-score of 0.92, while increasing the mAP of a YOLO detector from 0.52 to 0.91, representing a 75% improvement in detection performance with 35% less human intervention. This contribution advances the field of Data-Centric AI by introducing a teaching-inspired methodology that combines human advice and reinforcement learning to accelerate the creation of annotations in domains with scarce labeled data.
- 5 January 2026
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 39 (5) , 4145-4156
Enabling Autonomous Data Annotation in Mammography Image: A Human-in-the-Loop Reinforcement Learning Approach
The vast majority of work in computer vision focuses on developing and applying new machine learning models, which often rely on large amounts of labeled training data. However, annotation is costly and time-consuming. This paper presents an approach based on Deep Reinforcement Learning (DRL) to automatically generate new annotations and reduce human effort in the preparation of training data for supervised learning models in object detection. Our methodology introduces a virtual agent trained with human guidance, inspired by constructivist teaching methods, where interaction with a human teacher supports the agent's learning process. The proposed approach, named "Try a Little More" (TLM), employs active learning to identify uncertain cases and request human intervention during training, progressively enhancing the agent's ability to annotate autonomously. We evaluated our methodology on a mammography dataset, where the agent created bounding box annotations later used in a state-of-the-art supervised object detection algorithm. The results demonstrate that the proposed approach improves the quality and quantity of training data, surpassing existing reinforcement learning and human-computer interaction methods. Quantitatively, TLM generated 414 new annotations with an IoU of 0.86 and F1-score of 0.92, while increasing the mAP of a YOLO detector from 0.52 to 0.91, representing a 75% improvement in detection performance with 35% less human intervention. This contribution advances the field of Data-Centric AI by introducing a teaching-inspired methodology that combines human advice and reinforcement learning to accelerate the creation of annotations in domains with scarce labeled data.
- 11 December 2024
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 38 (4) , 2260-2280
Towards Automated Semantic Segmentation in Mammography Images for Enhanced Clinical Applications
Mammography images are widely used to detect non-palpable breast lesions or nodules, aiding in cancer prevention and enabling timely intervention when necessary. To support medical analysis, computer-aided detection systems can automate the segmentation of landmark structures, which is helpful in locating abnormalities and evaluating image acquisition adequacy. This paper presents a deep learning-based framework for segmenting the nipple, the pectoral muscle, the fibroglandular tissue, and the fatty tissue in standard-view mammography images. To the best of our knowledge, we introduce the largest dataset dedicated to mammography segmentation of key anatomical structures, specifically designed to train deep learning models for this task. Through comprehensive experiments, we evaluated various deep learning model architectures and training configurations, demonstrating robust segmentation performance across diverse and challenging cases. These results underscore the framework’s potential for clinical integration. In our experiments, four semantic segmentation architectures were compared, all showing suitability for the target problem, thereby offering flexibility in model selection. Beyond segmentation, we introduce a suite of applications derived from this framework to assist in clinical assessments. These include automating tasks such as multi-view lesion registration and anatomical position estimation, evaluating image acquisition quality, measuring breast density, and enhancing visualization of breast tissues, thus addressing critical needs in breast cancer screening and diagnosis.
- 11 December 2024
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 38 (4) , 2260-2280
Towards Automated Semantic Segmentation in Mammography Images for Enhanced Clinical Applications
Mammography images are widely used to detect non-palpable breast lesions or nodules, aiding in cancer prevention and enabling timely intervention when necessary. To support medical analysis, computer-aided detection systems can automate the segmentation of landmark structures, which is helpful in locating abnormalities and evaluating image acquisition adequacy. This paper presents a deep learning-based framework for segmenting the nipple, the pectoral muscle, the fibroglandular tissue, and the fatty tissue in standard-view mammography images. To the best of our knowledge, we introduce the largest dataset dedicated to mammography segmentation of key anatomical structures, specifically designed to train deep learning models for this task. Through comprehensive experiments, we evaluated various deep learning model architectures and training configurations, demonstrating robust segmentation performance across diverse and challenging cases. These results underscore the framework’s potential for clinical integration. In our experiments, four semantic segmentation architectures were compared, all showing suitability for the target problem, thereby offering flexibility in model selection. Beyond segmentation, we introduce a suite of applications derived from this framework to assist in clinical assessments. These include automating tasks such as multi-view lesion registration and anatomical position estimation, evaluating image acquisition quality, measuring breast density, and enhancing visualization of breast tissues, thus addressing critical needs in breast cancer screening and diagnosis.
- 1 December 2023
- Published by Springer Nature
- p. 56-67
A Data-Centric Approach for Pectoral Muscle Deep Learning Segmentation Enhancements in Mammography Images
- 1 December 2023
- Published by Springer Nature
- p. 56-67
A Data-Centric Approach for Pectoral Muscle Deep Learning Segmentation Enhancements in Mammography Images
- 10 July 2020
- Published by Springer Nature
- p. 385-397
A Self-adaptive Serious Game for Eye-Hand Coordination Training
- 10 July 2020
- Published by Springer Nature
- p. 385-397
A Self-adaptive Serious Game for Eye-Hand Coordination Training