Andrea Pinna
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Andrea Pinna has a total of 39 co-authors. Here are the top co-authors by number of shared publications:
- Tomasz Kryjak
AGH University of Krakow
5 shared publications - Bertrand Granado
Université Pierre et Marie Curie
4 shared publications - Kamil Jeziorek
Embedded Vision Systems Group AGH University of Krakow Poland
4 shared publications - Chuquimia Orlando
LIP6, CNRS UMR 7606, Sorbonne Université, Paris, France
3 shared publications - Krzysztof Blachut
AGH University of Krakow
3 shared publications - Piotr Wzorek
AGH University of Krakow
3 shared publications
Latest Publications
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 59 (11) , 144-148
The Application-Driven Paradigm: A Strategic Framework for European Cyber–Physical Systems Autonomy
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 59 (11) , 144-148
The Application-Driven Paradigm: A Strategic Framework for European Cyber–Physical Systems Autonomy
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 344-345
Activation Function Approximations Using a Reconfigurable Multilayer Perceptron
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 344-345
Activation Function Approximations Using a Reconfigurable Multilayer Perceptron
- Open Access
- 30 May 2026
- Published by Elsevier in Journal of Systems Architecture
- Vol. 177
Embedded graph convolutional networks for real-time event data processing on SoC FPGAs
- 30 May 2026
- Published by Elsevier in Journal of Systems Architecture
- Vol. 177
Open AccessEmbedded graph convolutional networks for real-time event data processing on SoC FPGAs
- Open Access
- 30 September 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Reconfigurable Technology and Systems
- Vol. 18 (3) , 1-22
VERSATILE: Very Fast Partial Reconfiguration Controller
Dynamically reconfigurable architectures allow sharing of hardware resources, which is particularly beneficial for small low-end FPGAs. Based on the online modification of parts of the circuit, these architectures require partial reconfiguration of the chip. Reducing resource availability or usage comes at the cost of a performance penalty affecting execution time. The challenge lies in bitstream management, especially for complex applications that often exceed the internal memory capacity of the FPGA (BRAM). Consequently, the time penalty arises from the need to retrieve partial bitstreams from external memory (e.g., often a DDR) each time it is necessary. Current state-of-the-art reconfiguration controllers are limited to a throughput of 400 MB/s, significantly penalizing reconfiguration times and making dynamic reconfiguration unattractive for real-life applications (e.g., video processing, machine learning applications, and continual and federated learning for embedded systems). This article introduces a novel partial reconfiguration controller architecture that achieves a throughput of up to 1.396 GB/s, a 3.49× acceleration over existing controllers. The reduced reconfiguration time allows the practical use of dynamic reconfiguration with fewer performance penalties. Additionally, the article compares various reconfiguration controllers in terms of time penalties and offers a tradeoff between algorithm complexity, FPGA resources, and performance.
- 30 September 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Reconfigurable Technology and Systems
- Vol. 18 (3) , 1-22
Open AccessVERSATILE: Very Fast Partial Reconfiguration Controller
Dynamically reconfigurable architectures allow sharing of hardware resources, which is particularly beneficial for small low-end FPGAs. Based on the online modification of parts of the circuit, these architectures require partial reconfiguration of the chip. Reducing resource availability or usage comes at the cost of a performance penalty affecting execution time. The challenge lies in bitstream management, especially for complex applications that often exceed the internal memory capacity of the FPGA (BRAM). Consequently, the time penalty arises from the need to retrieve partial bitstreams from external memory (e.g., often a DDR) each time it is necessary. Current state-of-the-art reconfiguration controllers are limited to a throughput of 400 MB/s, significantly penalizing reconfiguration times and making dynamic reconfiguration unattractive for real-life applications (e.g., video processing, machine learning applications, and continual and federated learning for embedded systems). This article introduces a novel partial reconfiguration controller architecture that achieves a throughput of up to 1.396 GB/s, a 3.49× acceleration over existing controllers. The reduced reconfiguration time allows the practical use of dynamic reconfiguration with fewer performance penalties. Additionally, the article compares various reconfiguration controllers in terms of time penalties and offers a tradeoff between algorithm complexity, FPGA resources, and performance.
- 1 April 2025
- Published by Springer Nature
- p. 57-68
Endoscopy Image Classification for Wireless Capsules with CNNs on Microcontroller-Based Platforms
- 1 April 2025
- Published by Springer Nature
- p. 57-68
Endoscopy Image Classification for Wireless Capsules with CNNs on Microcontroller-Based Platforms
- 21 October 2024
- Published by Elsevier in Microprocessors and Microsystems
- Vol. 111
Algorithms for scheduling CNNs on multicore MCUs at the neuron and layer levels
- 21 October 2024
- Published by Elsevier in Microprocessors and Microsystems
- Vol. 111
Algorithms for scheduling CNNs on multicore MCUs at the neuron and layer levels
- Open Access
- 26 September 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
- Vol. 24 (1) , 1-26
ZIP-CNN: Design Space Exploration for CNN Implementation within a MCU
Embedded systems based on Microcontroller Units (MCUs) often gather significant quantities of data and solve various issues. Convolutional Neural Networks (CNNs) have proven their effectiveness in solving computer vision and natural language processing tasks. However, implementing CNNs within MCUs is challenging due to their high inference costs, which varies widely depending on hardware targets and CNN topologies. Despite state-of-the-art advancements, no efficient design space exploration solutions handle the wide variety of implementation solutions. In this article, we introduce the ZIP-CNN design space exploration methodology, which facilitates CNN implementation within MCUs. We developed a model that quantitatively estimates the latency, energy consumption, and memory space required to run a CNN within an MCU. This model accounts for algorithmic reductions such as knowledge distillation, pruning, or quantization and applies to any CNN topology. To demonstrate the efficiency of our methodology, we investigated LeNet5, ResNet8, and ResNet26 within three different MCUs. We made materials and supplementary results available in a GitHub repository: https://github.com/ThGbay/ZIP-CNN . The proposed method was empirically verified on three hardware targets running at 14 different operating frequencies. The three CNN topologies investigated were implemented in their default configuration in FP32, and also reduced with INT8 quantization, pruning at five different rates and with knowledge distillation. The estimates of our model are very reliable with an error of 3.29% to 15.23% for latency, 3.12% to 10.34% for energy consumption, and 1.95% to 6.31% for memory space. These results are based on on-device measurements.
- 26 September 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
- Vol. 24 (1) , 1-26
Open AccessZIP-CNN: Design Space Exploration for CNN Implementation within a MCU
Embedded systems based on Microcontroller Units (MCUs) often gather significant quantities of data and solve various issues. Convolutional Neural Networks (CNNs) have proven their effectiveness in solving computer vision and natural language processing tasks. However, implementing CNNs within MCUs is challenging due to their high inference costs, which varies widely depending on hardware targets and CNN topologies. Despite state-of-the-art advancements, no efficient design space exploration solutions handle the wide variety of implementation solutions. In this article, we introduce the ZIP-CNN design space exploration methodology, which facilitates CNN implementation within MCUs. We developed a model that quantitatively estimates the latency, energy consumption, and memory space required to run a CNN within an MCU. This model accounts for algorithmic reductions such as knowledge distillation, pruning, or quantization and applies to any CNN topology. To demonstrate the efficiency of our methodology, we investigated LeNet5, ResNet8, and ResNet26 within three different MCUs. We made materials and supplementary results available in a GitHub repository: https://github.com/ThGbay/ZIP-CNN . The proposed method was empirically verified on three hardware targets running at 14 different operating frequencies. The three CNN topologies investigated were implemented in their default configuration in FP32, and also reduced with INT8 quantization, pruning at five different rates and with knowledge distillation. The estimates of our model are very reliable with an error of 3.29% to 15.23% for latency, 3.12% to 10.34% for energy consumption, and 1.95% to 6.31% for memory space. These results are based on on-device measurements.
- 22 June 2024
- Published by Springer Nature
- p. 110-122
Optimising Graph Representation for Hardware Implementation of Graph Convolutional Networks for Event-Based Vision
- 22 June 2024
- Published by Springer Nature
- p. 110-122
Optimising Graph Representation for Hardware Implementation of Graph Convolutional Networks for Event-Based Vision
- Open Access
Optimising Graph Representation for Hardware Implementation of Graph Convolutional Networks for Event-based Vision
Event-based vision is an emerging research field involving processing data generated by Dynamic Vision Sensors (neuromorphic cameras). One of the latest proposals in this area are Graph Convolutional Networks (GCNs), which allow to process events in its original sparse form while maintaining high detection and classification performance. In this paper, we present the hardware implementation of a~graph generation process from an event camera data stream, taking into account both the advantages and limitations of FPGAs. We propose various ways to simplify the graph representation and use scaling and quantisation of values. We consider both undirected and directed graphs that enable the use of PointNet convolution. The results obtained show that by appropriately modifying the graph representation, it is possible to create a~hardware module for graph generation. Moreover, the proposed modifications have no significant impact on object detection performance, only 0.08% mAP less for the base model and the N-Caltech data set.Finally, we describe the proposed hardware architecture of the graph generation module.
Open AccessOptimising Graph Representation for Hardware Implementation of Graph Convolutional Networks for Event-based Vision
Event-based vision is an emerging research field involving processing data generated by Dynamic Vision Sensors (neuromorphic cameras). One of the latest proposals in this area are Graph Convolutional Networks (GCNs), which allow to process events in its original sparse form while maintaining high detection and classification performance. In this paper, we present the hardware implementation of a~graph generation process from an event camera data stream, taking into account both the advantages and limitations of FPGAs. We propose various ways to simplify the graph representation and use scaling and quantisation of values. We consider both undirected and directed graphs that enable the use of PointNet convolution. The results obtained show that by appropriately modifying the graph representation, it is possible to create a~hardware module for graph generation. Moreover, the proposed modifications have no significant impact on object detection performance, only 0.08% mAP less for the base model and the N-Caltech data set.Finally, we describe the proposed hardware architecture of the graph generation module.
- Open Access
- 30 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
Memory-Efficient Graph Convolutional Networks for Object Classification and Detection with Event Cameras
Recent advances in event camera research emphasize processing data in its original sparse form, which allows the use of its unique features such as high temporal resolution, high dynamic range, low latency, and resistance to image blur. One promising approach for analyzing event data is through graph convolutional networks (GCNs). However, current research in this domain primarily focuses on optimizing computational costs, neglecting the associated memory costs. In this paper, we consider both factors together in order to achieve satisfying results and relatively low model complexity. For this purpose, we performed a comparative analysis of different graph convolution operations, considering factors such as execution time, the number of trainable model parameters, data format requirements, and training outcomes. Our results show a 450-fold reduction in the number of parameters for the feature extraction module and a 4.5-fold reduction in the size of the data representation while maintaining a classification accuracy of 52.3%, which is 6.3% higher compared to the operation used in state-of-the-art approaches. To further evaluate performance, we implemented the object detection architecture and evaluated its performance on the N-Caltech101 dataset. The results showed an accuracy of 53.7 % mAP@0.5 and reached an execution rate of 82 graphs per second.
- 30 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
Open AccessMemory-Efficient Graph Convolutional Networks for Object Classification and Detection with Event Cameras
Recent advances in event camera research emphasize processing data in its original sparse form, which allows the use of its unique features such as high temporal resolution, high dynamic range, low latency, and resistance to image blur. One promising approach for analyzing event data is through graph convolutional networks (GCNs). However, current research in this domain primarily focuses on optimizing computational costs, neglecting the associated memory costs. In this paper, we consider both factors together in order to achieve satisfying results and relatively low model complexity. For this purpose, we performed a comparative analysis of different graph convolution operations, considering factors such as execution time, the number of trainable model parameters, data format requirements, and training outcomes. Our results show a 450-fold reduction in the number of parameters for the feature extraction module and a 4.5-fold reduction in the size of the data representation while maintaining a classification accuracy of 52.3%, which is 6.3% higher compared to the operation used in state-of-the-art approaches. To further evaluate performance, we implemented the object detection architecture and evaluated its performance on the N-Caltech101 dataset. The results showed an accuracy of 53.7 % mAP@0.5 and reached an execution rate of 82 graphs per second.