Ultrasound Imaging ‒ LTS5 ‐ EPFL

LTS5 research in ultrasound imaging

Regularization methods in ultrasound imaging

In terms of acoustic wave propagation, ultrasound imaging poses an inverse problem which relates the tissue reflectivity (image under scrutiny) to the backscattered echoes (the measurements). Standard imaging methods are based on a well-known algorithm called Delay-and-sum (DAS). The major advantage of DAS is that it is fast and highly parallelizable. Its main drawback is that it is an inaccurate solution of the inverse problem which leads to low-quality images. One research topic in LTS5 consists in exploring alternatives to DAS mainly based on regularization approaches. The idea is to inject prior knowledge (sparsity of the image in a well-chosen model) about the tissue reflectivity in order to leverage the ill-posedness of the problem. The problem is recast as a convex optimization algorithm and solved using standard algorithms such as ADMM, primal-dual forward backward or FISTA.

Figure: Cyst image reconstructed with the DAS algorithm (left) and with the sparse regularization approach (right).

Compressed sensing and deep learning in ultrasound imaging

Compressed sensing or compressive sensing (CS) is a well-known mathematical framework which ensures perfect recovery of compressible signals from random projections with fewer samples than Nyquist rate. The recovery is achieved by solving a convex optimization problem where the most compressible solution which fits to the measurements is identified. This framework has been substantially studied in the last ten years and has lead to great successes such as: the Rice single pixel camera or the Sparse MRI. At LTS5, we are currently working on applying the CS framework to ultrasound imaging. The first idea is to exploit the intrinsic structure of US data which can be modelled as a stream of pulses, compressible in a redundant dictionary made of shifted replica of the ultrasound pulse. The second idea consists in designing acquisition schemes that are most suited to the CS framework. The last work focuses on designing efficient reconstruction algorithms. To do so, we are intensively working on deep learning algorithms which are becoming extremely popular nowadays due to their ability to perform a various of tasks (object recognition, classification, text generation, image colorization etc.). At LTS5, we are currently focusing on three main deep neural netwrok architectures for ultrasound imaging:

DNN that maps convex optimization algorithms such as Fast-Iterative thredsholding algorithm (FISTA);
Fully connected architectures (Stacked denoising autoencoders, Multi-layer perceptron);
Convolutional neural networks.

An Implantable Wireless System for Remote Hemodynamic Monitoring of Heart Failure Patients

M. Besirli; K. Ture; M. Beghetti; F. Maloberti; C. Dehollain et al.

Ieee Transactions On Biomedical Circuits And Systems. 2023-08-01. Vol. 17, num. 4, p. 688-700. DOI : 10.1109/TBCAS.2023.3273711.

LTS5 research in ultrasound imaging

An Implantable Wireless System for Remote Hemodynamic Monitoring of Heart Failure Patients

Design Space Exploration for Partitioning Dataflow Program on CPU-GPU Heterogeneous System

FM -Directories: Extending the Burrows-Wheeler Transform for String Labeled Vertex Graphs of (Almost) Arbitrary Topology

Integrated Electronics for Deep Implants to Remotely Monitor Hemodynamics

Toward a common standard for data and specimen provenance in life sciences

Compilation and Design Space Exploration of Dataflow Programs for Heterogeneous CPU-GPU Platforms

The impact of NFT profile pictures within social network communities

An Energy-Efficient Bridge-to-Digital Converter for Implantable Pressure Monitoring Systems

An Implantable Inductive Sensor for Direct and Continuous Monitoring of the Pulmonary Artery Cross-Sectional Area

A Capacitively-Coupled Chopper Instrumentation Amplifier for Implantable Bridge Sensor Systems

Dynamic SIMD Parallel Execution on GPU from High-Level Dataflow Synthesis

Performance Estimation of High-Level Dataflow Program on Heterogeneous Platforms by Dynamic Network Execution

MPEG-G Reference-Based Compression of Unaligned Reads Through Ultra-Fast Alignments

Inter-actions parallel execution on GPU from high-level dataflow synthesis

SIMD Parallel Execution on GPU from High-Level Dataflow Synthesis

Performance Estimation of High-Level Dataflow Program on Heterogeneous Platforms

Networks of Ethereum Non-Fungible Tokens: A graph-based analysis of the ERC-721 ecosystem

A 0.4 nJ Excitation Energy Bridge-to-Digital Converter for Implantable Pulmonary Artery Pressure Monitoring

Methodologies for Synthesizing and Analyzing Dynamic Dataflow Programs in Heterogeneous Systems for Edge Computing

An Introduction to MPEG-G: The First Open ISO/IEC Standard for the Compression and Exchange of Genomic Sequencing Data

CMOS-Based Readout and Control Electronics for Microgrippers

Programming Heterogeneous CPU-GPU Systems by High-Level Dataflow Synthesis

Pipeline Synthesis and Optimization from Branched Feedback Dataflow Programs

Fully Soft-Switched High Step-Up Nonisolated Three-Port DC-DC Converter Using GaN HEMTs

High Precision Capacitive Moisture Sensor for Polymers: Modeling and Experiments

Low-Cost Readout Electronics for Piezoresistive MEMS-Based Transducers

On the relevance of quality score metadata in genomic sequence data for omics applications

A CMOS Analog Front-End for Implantable Pulmonary Artery Pressure Monitoring System

An Heterogeneous Compiler Of Dataflow Programs For Zynq Platforms

Modeling Dielectric Constant Variability in Aggregate Polymers from CV Measurements

Composite Data Types in Dynamic Dataflow Languages as Copyless Memory Sharing Mechanism

Toward a Dynamic Threshold for Quality-Score Distortion in Reference-Based Alignment

Shared-variable Synchronization Approaches for Dynamic Data Flow Programs

Efficient Dynamic Optimisation Heuristics for Dataflow Pipelines

High Precision Capacitive Moisture Sensor for Polymers

Execution Trace Graph of Dataflow Process Networks

Lossy compression of quality scores in differential gene expression: A first assessment and impact analysis

Transcriptome reconstruction with quality score distortion in reference-based alignment

High-Precision Performance Estimation for the Design Space Exploration of Dynamic Dataflow Programs

Design space exploration of dataflow-based Smith-Waterman FPGA implementations

Buffer dimensioning for throughput improvement of dynamic dataflow signal processing applications on multi-core platforms

High level synthesis of Smith-Waterman dataflow implementations

Performance estimation of program partitions on multi-core platforms, 2016 26th International Workshop on Power and Timing Modeling, Optimization and Simulation (PATMOS)

MPEG-G the emerging standard for genomic data compression

Differential gene expression with lossy compression of quality scores in RNA-seq data

Systematic Design Space Exploration of Dynamic Dataflow Programs for Multi-core Platforms

Clock-Gating of Streaming Applications for Energy Efficient Implementations on FPGAs

Trace-based manycore partitioning of stream-processing applications

High-level system synthesis and optimization of dataflow programs for MPSoCs

High-level synthesis of dynamic dataflow programs on heterogeneous MPSoC platforms

Design Space Exploration Problem Formulation for Dataflow Programs on Heterogeneous Architectures

Programming Models and Methods for Heterogeneous Parallel Embedded Systems

High-Precision Performance Estimation of Dynamic Dataflow Programs

On the Development and Optimization of HEVC Video Decoders Using High-Level Dataflow Modeling

Comparison of high-throughput sequencing data compression tools

An Evaluation Framework for Lossy Compression of Genome Sequencing Quality Values

Performance Estimation Based Multicriteria Partitioning Approach for Dynamic Dataflow Programs

Dataflow Programs Analysis and Optimization Using Model Predictive Control Techniques. Two Examples of Bounded Buffer Scheduling: Deadlock Avoidance and Deadlock Recovery Strategies

Automated Design Flow for Multi-Functional Dataflow-Based Platforms

A Methodology for Profiling and Partitioning Stream Programs on Many-core Architectures

Execution Trace Graph Based Multi-criteria Partitioning of Stream Programs

Synthesis and Optimization of Pipelines for HW Implementations of Dataflow Programs

Analysis and optimization of dynamic dataflow programs

Actor Merging for Dataflow Process Networks

High-level synthesis of dataflow programs for heterogeneous platforms

MPEG high efficient video coding stream programming and many-cores scalability

TURNUS: An open-source design space exploration framework for dynamic stream programs

Execution trace graph analysis of dataflow programs: Bounded buffer scheduling and deadlock recovery using model predictive control

Characterizing communication behavior of dataflow programs using trace analysis

Exploring MPEG HEVC decoder parallelism for the efficient porting onto many-core platforms

Dataflow programs analysis and optimization using model predictive control techniques: An example of bounded buffer scheduling

Automated design flow for coarse-grained reconfigurable platforms: An RVC-CAL multi-standard decoder use-case

Coarse Grain Clock Gating Of Streaming Applications In Programmable Logic Implementations

A Methodology For Optimizing Buffer Sizes Of Dynamic Dataflow Fpgas Implementations

ECMA-407: A New 3D audio codec implementation up to NHK 22.2

ECMA-407: New Approaches to 3D Audio Content Data Rate Reduction with RVC-CAL

Optimizing Dataflow Programs for Hardware Synthesis

High-level dataflow design of signal processing systems for reconfigurable and multicore heterogeneous platforms

Automated Qoe Evaluation Of Dynamic Adaptive Streaming Over Http