Cristian Sestito

Dr. Cristian Sestito

Department of Informatics, Modeling, Electronics and System Engineering, University of Calabria, Rende, Italy
orcid

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
6 September 2024 in Science Advances
Verified
Journal Article
Closed access
CSCaterina Sbandati
Caterina Sbandati
SSSpyros Stathopoulos
Spyros Stathopoulos
PFPatrick Foster
Patrick Foster
NPNoam D. Peer
Noam D. Peer
Cristian SestitoCristian Sestito
Cristian Sestito
Themis ProdromakisThemis Prodromakis
Themis Prodromakis
Implantable devices hold the potential to address conditions currently lacking effective treatments, such as drug-resistant neural impairments and prosthetic control. Medical devices need to be biologically compatible while providing enhanced performance metrics of low-power consumption, high accuracy, small size, and minimal latency to enable ongoing intervention in brain function. Here, we demonstrate a memristor-based processing system for single-trial detection of behaviorally meaningful brain signals within a timeframe that supports real-time closed-loop intervention. We record neural activity from the reward center of the brain, the ventral tegmental area, in rats trained to associate a musical tone with a reward, and we use the memristors built-in thresholding properties to detect nontrivial biomarkers in local field potentials. This approach yields consistent and accurate detection of biomarkers >98% while maintaining power consumption as low as 4.14 nanowatt per channel. The efficacy of our system’s capabilities to process real-time in vivo neural data paves the way for low-power chronic neural activity monitoring and biomedical implants.

12 October 2021 in Journal of Imaging
Verified
Journal Article
Open access
PDF file
Cristian SestitoCristian Sestito
Cristian Sestito
FSFanny Spagnolo
Fanny Spagnolo
SPStefania Perri
Stefania Perri
Nowadays, computer vision relies heavily on convolutional neural networks (CNNs) to perform complex and accurate tasks. Among them, super-resolution CNNs represent a meaningful example, due to the presence of both convolutional (CONV) and transposed convolutional (TCONV) layers. While the former exploit multiply-and-accumulate (MAC) operations to extract features of interest from incoming feature maps (fmaps), the latter perform MACs to tune the spatial resolution of the received fmaps properly. The ever-growing real-time and low-power requirements of modern computer vision applications represent a stimulus for the research community to investigate the deployment of CNNs on well-suited hardware platforms, such as field programmable gate arrays (FPGAs). FPGAs are widely recognized as valid candidates for trading off computational speed and power consumption, thanks to their flexibility and their capability to also deal with computationally intensive models. In order to reduce the number of operations to be performed, this paper presents a novel hardware-oriented algorithm able to efficiently accelerate both CONVs and TCONVs. The proposed strategy was validated by employing it within a reconfigurable hardware accelerator purposely designed to adapt itself to different operating modes set at run-time. When characterized using the Xilinx XC7K410T FPGA device, the proposed accelerator achieved a throughput of up to 2022.2 GOPS and, in comparison to state-of-the-art competitors, it reached an energy efficiency up to 2.3 times higher, without compromising the overall accuracy.

4 August 2023 in Journal of Signal Processing Systems
Verified
Journal Article
Open access
PDF file
Cristian SestitoCristian Sestito
Cristian Sestito
SPStefania Perri
Stefania Perri
RSRobert Stewart
Robert Stewart
Deep Learning (DL) is pervasive across a wide variety of domains. Convolutional Neural Networks (CNNs) are often used for image processing DL applications. Modern CNN models are growing to meet the needs of more sophisticated tasks, e.g. using Transposed Convolutions (TCONVs) for image decompression and image generation. Such state-of-the-art DL models often target GPU-based high-performance architectures, due to the high computational and hardware resource needs of TCONV layers. To avoid prohibitive GPU energy costs, CNNs are increasingly deployed to decentralized embedded autonomous devices, such as Field Programmable Gate Arrays (FPGAs). However, this poses challenges for designing efficient hardware implementations of TCONV layers. This paper presents a parameterized design and implementation of a new TCONV module, which is synthesizable onto FPGAs. It is implemented using the High-Level Synthesis (HLS), through a C++ template to parameterize its functional and non-functional properties. These parameters allow kernel sizes, image sizes, quantization and parallelism to be varied by users. With a systematic exploration in this design space, we find an optimal instance of this TCONV module that achieves 6.25 Giga Outputs per Second (Gout/s) using just 1.53 W of power. We then use our TCONV layer in two neural networks for image decompression and image generation. Image decompression achieves a speed throughput of more than 30K frames-per-second (fps) using only the 16% of resources on average, image generation achieves an energy efficiency of 324 fps/W and outperforms comparable state-of-the-art models by at least 7.3×.

19 November 2024 in Electronics
Verified
Journal Article
Open access
PDF file
SSSimona Sibio
Simona Sibio
Cristian SestitoCristian Sestito
Cristian Sestito
SSSouheil Ben Smida
Souheil Ben Smida
YDYuan Ding
Yuan Ding
GGGeorge Goussetis
George Goussetis
This paper presents a deep learning algorithm for channel estimation in 5G New Radio (NR). The classical approach that uses neural networks for channel estimation requires more than one stage to obtain the full channel matrix. First, the channel has to be constructed by the received reference signal, and then, the precision is improved. In contrast, to reduce the computational cost, the proposed neural network method generates the channel matrix from the information captured from a few subcarriers along the slot. This information is extrapolated by applying the Least Square technique only on the Demodulation Reference Signal (DMRS). The received DMRS placed in the grid can be seen as a 2D low-resolution image and it is processed to generate the full channel matrix. To reduce complexity in the hardware implementation, the convolutional neural network (CNN) structure is selected. This solution is analyzed comparing the Mean Square Error (MSE) and the computational cost with other deep learning-based channel estimation, as well as the traditional channel estimation methods. It is demonstrated that the proposed neural network delivers substantial complexity savings and favorable error performance. It reduces the computational cost by an order of magnitude, and it has a maximum error discrepancy of 0.018 at 5 dB compared to Minimum Mean Square Error (MMSE) channel estimation.

1 January 2025 in IEEE Transactions on Circuits and Systems I: Regular Papers
Verified
Journal Article
Closed access
Cristian SestitoCristian Sestito
Cristian Sestito
SAShady Agwa
Shady Agwa
Themis ProdromakisThemis Prodromakis
Themis Prodromakis
Modern hardware architectures for Convolutional Neural Networks (CNNs), other than targeting high performance, aim at dissipating limited energy. Reducing the data movement cost between the computing cores and the memory is a way to mitigate the energy consumption. Systolic arrays are suitable architectures to achieve this objective: they use multiple processing elements that communicate each other to maximize data utilization, based on proper dataflows like the weight stationary and row stationary. Motivated by this, we have proposed TrIM, an innovative dataflow based on a triangular movement of inputs, and capable to reduce the number of memory accesses by one order of magnitude when compared to state-of-the-art systolic arrays. In this paper, we present a TrIM-based hardware architecture for CNNs. As a showcase, the accelerator is implemented onto a Field Programmable Gate Array (FPGA) to execute the VGG-16 and AlexNet CNNs. The architecture achieves a peak throughput of 453.6 Giga Operations per Second, outperforming a state-of-the-art row stationary systolic array up to $\sim 3 \times$ in terms of memory accesses, and being up to $ \sim 11.9 \times$ more energy-efficient than other FPGA accelerators.

25 August 2020 in Journal of Imaging
Verified
Journal Article
Open access
PDF file
SPStefania Perri
Stefania Perri
Cristian SestitoCristian Sestito
Cristian Sestito
FSFanny Spagnolo
Fanny Spagnolo
Pasquale CorsonelloPasquale Corsonello
Pasquale Corsonello
Today, convolutional and deconvolutional neural network models are exceptionally popular thanks to the impressive accuracies they have been proven in several computer-vision applications. To speed up the overall tasks of these neural networks, purpose-designed accelerators are highly desirable. Unfortunately, the high computational complexity and the huge memory demand make the design of efficient hardware architectures, as well as their deployment in resource- and power-constrained embedded systems, still quite challenging. This paper presents a novel purpose-designed hardware accelerator to perform 2D deconvolutions. The proposed structure applies a hardware-oriented computational approach that overcomes the issues of traditional deconvolution methods, and it is suitable for being implemented within any virtually system-on-chip based on field-programmable gate array devices. In fact, the novel accelerator is simply scalable to comply with resources available within both high- and low-end devices by adequately scaling the adopted parallelism. As an example, when exploited to accelerate the Deep Convolutional Generative Adversarial Network model, the novel accelerator, running as a standalone unit implemented within the Xilinx Zynq XC7Z020 System-on-Chip (SoC) device, performs up to 72 GOPs. Moreover, it dissipates less than 500mW@200MHz and occupies 5.6%, 4.1%, 17%, and 96%, respectively, of the look-up tables, flip-flops, random access memory, and digital signal processors available on-chip. When accommodated within the same device, the whole embedded system equipped with the novel accelerator performs up to 54 GOPs and dissipates less than 1.8W@150MHz. Thanks to the increased parallelism exploitable, more than 900 GOPs can be executed when the high-end Virtex-7 XC7VX690T device is used as the implementation platform. Moreover, in comparison with state-of-the-art competitors implemented within the Zynq XC7Z045 device, the system proposed here reaches a computational capability up to 20% higher, and saves more than 60% and 80% of power consumption and logic resources requirement, respectively, using 5.7× fewer on-chip memory resources.