The field of computer vision deals with the acquisition, processing, and analysis of digital images in order to develop applications for object detection, segmentation, and vision based learning tasks. The objective is to extract information from images and videos, which may come from various sensors (such as multiple cameras, medical scanners), using traditional and deep learning techniques. Studied topics include: scene reconstruction, event detection, video tracking, object recognition, writing analysis, and image restoration.

• Image processing and 3D modeling of historical musical instruments: in progress since 2014, in collaboration with Arvedi Laboratory of Non Invasive Diagnostics of University of Pavia. The objective is the analysis of surface and morfology of historical musical instruments, with a special focus on violins and other string musical instruments held in the Museo del Violino of Cremona (Italy).
• Analysis of frescoes: study of methods for the reconstruction of damaged frescoes or, more in general, fragmented images. In particular: creation of a dataset of simulated fresco fragments in collaboration with the Université Paris-Saclay and the Università degli Studi di Salerno (project ARTEAK, from September 2021 to August 2025); a creation of a dataset for digital anastylosis of frescoes (DAFNE, 2019); the application of generative AI techniques to virtual reconstruction of frescoes.
• Structural damage detection: adding artificial damage on 3D models obtained from photogrammetry; rendering of a semi-synthetic dataset to train a neural network; damage detection on real videos acquired post-earthquake. In collaboration with Eucentre (European Centre for Training and Research in Earthquake Engineering), the research contract TeamAware (from September 2021 to February 2024) is on the automatic visual identification of major injuries in buildings and infrastructure.
• Fall detection: we have used smart wearable devices and deep learning on embedded to detect accidental falls with recurrent neural networks. Datasets with simulated falls by volunteers have been collected: seven carry positions, seventeen different activities, forty volunteers, over five thousands tracks. Manual annotations on videos, basic for training, have been done.
• Collaborative robotics: deep reinforcement learning for collaborative robotics, focused on incremental autonomous learning, experience transfer, and robust avoidance strategy, with the virtualization of a real‐world robot to learn reaching a target while avoiding obstacles in a simulation environment.

Furthermore, in the past:

• Automatic screening of digital slides: the research was focused on automatic screening of digital slides and automatic segmentation of injuries (usually the diagnosis is made through a visual analysis of slides containing sentinel lymph node biopsy), applied in particular to the diagnosis of breast cancer metastases in lymph nodes.
• Writing analysis of Stradiviari's relics: carried out from 2016 to 2018, in collaboration with Arvedi Laboratory of Non Invasive Diagnostics of University of Pavia. The goal was studying the working notes present on Antonio Stardivari's relics, a collection of artifacts, technical drawings and wood molds used by the famous violin maker between the 17th and the 18th century, currently held in "Museo del Violino" of Cremona (Italy).
• The technology of a vehicle-based mobile mapping system has been applied to maintain an updated transportation database for road and railway inventory, for recognition of signs and tunnel inspection.
• Approaches and strategies of pattern recognition techniques have been used to improve aspects in the field of bioinformatics and computational biology, for modeling and comparing proteins 3D structures in Proteomics.
   In this context, in 2012, our laboratory has contributed to the TCBionformatics - GIRPR Technical Committee.
• Hierarchical architectures have been studied, developing the Papia, a processor array able to reconfigure itself as a pyramid, to respond to the heavy computational task of image processing.

---

Our latest Journal publications on Computer Vision

• Alireza Rezaei, Emanuel Aldea, Piercarlo Dondi, Sylvie Le Hégarat-Mascle, Marco Malagodi (2023). Multi-temporal image analysis for preventive conservation of historical musical instruments, ACM Journal on Computing and Cultural Heritage, Volume 16, Issue 2 Article No.: 35, pp. 1–19, DOI:10.1145/3575866.
• Alireza Rezaei, Sylvie Le Hégarat-Mascle, Emanuel Aldea, Piercarlo Dondi, Marco Malagodi (2022). A-contrario framework for detection of alterations in varnished surfaces, Journal of Visual Communication and Image Representation, Volume 83, February 2022, DOI:10.1016/j.jvcir.2021.103357.
• Giacomo Fiocco, Sebastian Gonzalez, Claudia Invernizzi, Tommaso Rovetta, Michela Albano, Piercarlo Dondi, Maurizio Licchelli, Fabio Antonacci, and Marco Malagodi (2021). Compositional and Morphological Comparison among Three Coeval Violins Made by Giuseppe Guarneri “del Gesù” in 1734, Coatings 2021, 11(8), 884, DOI:10.3390/coatings11080884.
• Virginio Cantoni, Gian Luca Foresti, Nicu Sebe (2021). Editorial for the Special Issue on the DAFNE Project (Digital Anastylosis of Frescoes challeNgE), Pattern Recognition Letters.
• Sangiovanni Bianca, Incremona Gian Paolo, Piastra Marco, Ferrara Antonella (2021). Self-Configuring Robot Path Planning with Obstacle Avoidance via Deep Reinforcement Learning, in IEEE Control Systems Letters, 5(2), pp. 397–402, DOI:10.1109/LCSYS.2020.3002852.