Learning features from visual content under limited supervision using multiple domains

• Grant number: 18/22482-0
• Support Opportunities: Regular Research Grants
• Start date: March 01, 2019
• End date: May 31, 2021
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Moacir Antonelli Ponti
• Grantee: Moacir Antonelli Ponti
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Associated researchers: John Collomosse
• Associated research grant(s): 19/16379-5 - Cross-domain visual representations under limited supervision and via transfer of learning, AP.R SPRINT

Abstract

Feature learning methods have reached state of the art on many applications, in particular in single-domain data, but also showing remarkable results for cross-domain datasets. Since collecting and labelling data can be costly and sometimes even not possible, it is paramount to investigate methods that can work with limited or no supervision. In this project we deal with the problem of feature learning from signal, image and video data under limited supervision. We address both the problem of finding a feature embedding for some given data, but also cross-domain matching, which is to find strategies to match content from a given task across different datasets or domains. Contributions towards investigating novel models and alternative architectures with respect to the methods often employed in the literature, including generative models, auto-encoders and others, which would allow to overcome the current challenges. (AU)

Scientific publications

(9)

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

BET, PATRICIA; CASTRO, PAULA C.; PONTI, MOACIR A.. Fall detection and fall risk assessment in older person using wearable sensors: A systematic review. INTERNATIONAL JOURNAL OF MEDICAL INFORMATICS, v. 130, OCT 2019. (13/07375-0, 18/22482-0)

DOS SANTOS, FERNANDO P.; RIBEIRO, LEONARDO S. F.; PONTI, MOACIR A.. Generalization of feature embeddings transferred from different video anomaly detection domains. JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION, v. 60, p. 407-416, APR 2019. (13/07375-0, 18/22482-0, 17/22366-8)

SOARES, V. H. A.; PONTI, M. A.; GONCALVES, R. A.; CAMPELLO, R. J. G. B.. Cattle counting in the wild with geolocated aerial images in large pasture areas. COMPUTERS AND ELECTRONICS IN AGRICULTURE, v. 189, OCT 2021. (18/22482-0)

BET, PATRICIA; CASTRO, PAULA C.; PONTI, MOACIR A.. Foreseeing future falls with accelerometer features in active community-dwelling older persons with no recent history of falls. Experimental Gerontology, v. 143, JAN 2021. (13/07375-0, 18/22482-0)

DOS SANTOS, FERNANDO P.; ZOR, CEMRE; KITTLER, JOSEF; PONTI, MOACIR A.. Learning image features with fewer labels using a semi-supervised deep convolutional network. NEURAL NETWORKS, v. 132, p. 131-143, DEC 2020. (19/07316-0, 18/22482-0)

VOGADO, LUIS; VERAS, RODRIGO; AIRES, KELSON; ARAUJO, FLAVIO; SILVA, ROMUERE; PONTI, MOACIR; TAVARES, JOAO MANUEL R. S.. Diagnosis of Leukaemia in Blood Slides Based on a Fine-Tuned and Highly Generalisable Deep Learning Model. SENSORS, v. 21, n. 9, MAY 2021. (18/22482-0)

OLIVEIRA DE RESENDE, DAMARES CRYSTINA; PONTI, MOACIR ANTONELLI. obust image features for classification and zero-shot tasks by merging visual and semantic attribute. NEURAL COMPUTING & APPLICATIONS, v. 34, n. 6, JAN 2022. (18/22482-0, 19/07316-0)

DOS SANTOS, FERNANDO PEREIRA; PONTI, MOACIR ANTONELLI; IEEE. Alignment of Local and Global Features from Multiple Layers of Convolutional Neural Network for Image Classification. 2019 32ND SIBGRAPI CONFERENCE ON GRAPHICS, PATTERNS AND IMAGES (SIBGRAPI), v. N/A, p. 8-pg., 2019-01-01. (13/07375-0, 18/22482-0)

NAZARE, TIAGO S.; DE MELLO, RODRIGO F.; PONTI, MOACIR A.; FARINELLA, GM; RADEVA, P; BRAZ, J; BOUATOUCH, K. Investigating 3D Convolutional Layers as Feature Extractors for Anomaly Detection Systems Applied to Surveillance Videos. VISAPP: PROCEEDINGS OF THE 16TH INTERNATIONAL JOINT CONFERENCE ON COMPUTER VISION, IMAGING AND COMPUTER GRAPHICS THEORY AND APPLICATIONS - VOL. 5: VISAPP, v. N/A, p. 10-pg., 2021-01-01. (13/07375-0, 15/04883-0, 18/22482-0)