Publications
Research contributions across machine learning, artificial intelligence, and related fields
Publications by Year
2025 (1)
A Benchmark Study of Deep Reinforcement Learning Algorithms for the Container Stowage Planning Problem
Yunqi Huang, Nishith Chennakeshava, Alexis Carras, Vladislav Neverov, Wei Liu, Aske Plaat, Yingjie Fan
This work evaluates five reinforcement learning algorithms for container stowage planning in maritime logistics. We developed a Gym environment incorporating crane scheduling and tested DQN, QR-DQN, A2C, PPO, and TRPO across scenarios of varying complexity to benchmark performance.
2024 (1)
An Adversarial Learning Approach to Generate Pressure Support Ventilation Waveforms for Asynchrony Detection
Liming Hao, Tom H.G.F. Bakkes, Anouk van Diepen, Nishith Chennakeshava, R. Arthur Bouwman, Ashley J.R. De Bie Dekker, Pierre H. Woerlee, Francesco Mojoli, Massimo Mischi, Yan Shi, Simona Turco
An adversarial learning approach to generate realistic pressure support ventilation waveforms for training asynchrony detection systems in mechanical ventilation.
2023 (2)
Ultrasound Signal Processing: From Models to Deep Learning
Ben Luijten, Nishith Chennakeshava, Yonina C. Eldar, Massimo Mischi, Ruud J.G. van Sloun
A comprehensive survey examining model-based deep learning approaches for medical ultrasound imaging that combine data-driven techniques with domain knowledge to improve reconstruction quality.
Shear-Wave Particle-Velocity Estimation and Enhancement Using a Multi-Resolution Convolutional Neural Network
Xingran Chen, Nishith Chennakeshava, Ruud R. Wildeboer, Massimo Mischi, Ruud J.G. van Sloun
A multi-resolution convolutional neural network approach for estimating and enhancing shear-wave particle velocity in ultrasound elastography imaging.
2022 (2)
Accelerated Intravascular Ultrasound Imaging using Deep Reinforcement Learning
Tristan S.W. Stevens, Nishith Chennakeshava, Frederik J. de Bruijn, Martin Pekař, Ruud J.G. van Sloun
This paper applies deep reinforcement learning with actor-critic methods to optimize adaptive acquisition policies for intravascular ultrasound, enabling faster imaging without compromising quality.
Deep Proximal Unfolding For Image Recovery from Under-Sampled Channel Data in Intravascular Ultrasound
Nishith Chennakeshava, Tristan S.W. Stevens, Frederik J. de Bruijn, Andrew Hancock, Martin Pekař, Yonina C. Eldar, Massimo Mischi, Ruud J.G. van Sloun
Deep proximal unfolding approach for recovering high-quality intravascular ultrasound images from under-sampled channel data.
2021 (1)
Deep Proximal Learning for High-Resolution Plane Wave Compounding
Nishith Chennakeshava, Ben Luijten, Massimo Mischi, Yonina C. Eldar, Ruud J.G. van Sloun
A physics-informed neural network approach that formulates plane wave compounding as an inverse problem, achieving high-resolution ultrasound images from just three transmissions using unfolded proximal gradient descent.
2020 (2)
Deep Learning for Classification and Localization of COVID-19 Markers in Point-of-Care Lung Ultrasound
Subhankar Roy, Willi Menapace, Sebastiaan Oei, Ben Luijten, Enrico Fini, Cristiano Saltori, Iris Huijben, Nishith Chennakeshava, Federico Mento, Alessandro Sentelli, Emanuele Peschiera, Riccardo Trevisan, Giovanni Maschietto, Elena Torri, Riccardo Inchingolo, Andrea Smargiassi, Gino Soldati, Paolo Rota, Andrea Passerini, Ruud J.G. van Sloun, Elisa Ricci, Libertario Demi
This paper presents deep learning approaches for analyzing lung ultrasound images to assist COVID-19 diagnosis. We introduce a novel fully-annotated dataset of lung ultrasound images collected from several Italian hospitals, with labels indicating the degree of disease severity at multiple levels. Key contributions include a novel network derived from Spatial Transformer Networks for simultaneous severity prediction and artifact localization, and a frame aggregation method using uninorms.
High Resolution Plane Wave Compounding Through Deep Proximal Learning
Nishith Chennakeshava, Ben Luijten, Oded Drori, Massimo Mischi, Yonina C. Eldar, Ruud J.G. van Sloun
Ultra-fast ultrasound imaging relies on coherent plane wave compounding to obtain sufficient spatial resolution and contrast. However, the process of coherent plane wave compounding incurs a loss in temporal resolution. This method achieves high-resolution results with fewer transmissions by embedding signal processing principles into the network design, demonstrating approximately an 8.2% improvement in PSNR over the next best network.