1,721,049 research outputs found
Design Candidate Identification via Kansei-VR & AHP approaches
No full textIn this work the Authors show the first results of a research activity aiming at the identification of the most appealing design candidate via a new integrated Kansei Engineering process. The target was achieved by means of immersive experiments performed in Virtual Reality (VR) along with an Analytic Hierarchy Process (AHP) performed in a visual desktop environment (PC). Both the approaches aim at the direct involvement of users into the design process, as early as possible.
Focusing on the synthesis phase, once implemented the design candidates by different technical features according to a Fractional Factorial Design, the concepts are evaluated by users. The data collected by asking users to judge them are analyzed via suitable methods to guarantee the above assessment. For this purpose, two different evaluation approaches, although at different stages of the design process, are tested: the first one relies on the user experience with the product in VR whereas the second is allowable for a much cheaper visual pairwise comparison in a PC-based experimental set-up. The original result is that the two approaches can be complementary rather than alternative; here is introduced the way to harmonize them in an integrated Kansei Engineering process, in order to improve and speed-up the synthesis phase.
To describe the two approaches and highlight their peculiarities, an application to the design of railway coach arrangement and furniture (briefly referred to as “train interior”) is presented
Dynamic Range-Invariant GAN Reconstruction via Optimized Target Training in Medical Ultrasound Imaging
Full text linkUltrasound imaging is sensitive to operatordependent
parameters such as dynamic range (DR), when
considering 8-bit reconstructed images, which can compromise
both clinical interpretation and the reliability of artificial
intelligence (AI)-based reconstruction pipelines. This work
validates an automatic dynamic range optimization (autoDR)
method as a standardized training reference for Generative
Adversarial Network (GAN) models. By evaluating GAN
robustness to DR variability, we demonstrate that autoDR
enables consistent, high-quality reconstructions across diverse
acquisition settings, outperforming classical enhancement
techniques. These findings highlight autoDR as a practical
solution for reducing operator dependence, improving
reproducibility, and a robust reference standard for GAN-based
ultrasound image reconstruction when there is no access to raw
radiofrequency data
High-frame-rate coherence imaging of the heart with ultrasound diverging waves
No full textSeveral techniques have been proposed up to now to achieve higher temporal resolution in echocardiography. Among these, the use of diverging beams, which insonify a large region of interest, allows to significantly increase the frame-rate, but at the cost of a reduced signal-to-noise ratio. For this reason, in this paper we propose to combine high-frame-rate imaging, by transmitting diverging waves (DWs), to the Short-Lag Spatial Coherence (SLSC) technique in reception, which provides images of the coherence of backscattered echoes and is known to yield improved contrast in scenarios with high-clutter. We test this combined method first on phantom acquisitions and then on in vivo cardiac scans, i.e. on apical views of the heart. Results show that SLSC can provide improved contrast ratio (CR) and generalized contrast-to-noise ratio (GCNR) with respect to the classic Delay and Sum (DAS) as the number of transmitted DWs increases, particularly when clutter is present. Indeed, cardiac images show improved apex visibility and artifact suppression in the heart chambers with SLSC, achieving high contrast and high frame-rate at the same time
Ultrasound Image Beamforming Optimization Using a Generative Adversarial Network
Full text linkRecently, research has been focusing on the development of artificial intelligence ultrasound beamforming methods to improve the contrast and resolution of B-mode images. In this work, we propose an innovative beamforming domain transfer method using a generative adversarial network (GAN). The GAN takes as input a plane-wave (PW) delay and sum (DAS) image and generates an image as if it had been acquired using the focused modality and reconstructed with the filtered Delay Multiply and Sum (F-DMAS) beamforming technique. A Verasonics Vantage 256 system (L11-5v linear array) was used to acquire 560 (480 and 80 for train and test set, respectively) in-vivo musculoskeletal US images. Images were acquired on five muscles (gastrocnemius lateralis, gastrocnemius medialis, vastus lateralis, vastus medialis, and biceps) on both sides of 14 healthy volunteers (50% female). RF data were acquired both in plane-wave (PW) and focused mode and beamformed using the UltraSound ToolBox (USTB). The DAS beamforming method was employed for PW data, whereas the focused data were reconstructed using F-DMAS. Various dynamic ranges (dR) were employed to create the final 8-bit PW DAS images (dR = 55, 65, 75, 85 dB) while an automatic dR was employed to optimize focused F-DMAS images. A Pix2Pix GAN architecture was designed to formulate the task of beamforming as the translation from one domain (PW DAS image) to another (focused F-DMAS image). Our GAN employed a UNet as the generator and a 3-layer fully convolutional PatchGAN as the discriminator. The proposed GAN architecture shows promising results, generating a GAN image comparable to the F-DMAS image, i.e., in terms of SSIM (0.5183 +/- 0.0437 and 0.5152 +/- 0.0519 for GAN images vs DAS images and F-DMAS images vs DAS images). Overall, our GAN enhances image quality and simulates focused F-DMAS beamforming starting from a PW DAS image without needing to access the raw RF data, which is typically unavailable with clinical ultrasound devices
Spatial Coherence Beamforming with Multi-Line Transmission to Enhance the Contrast of Coherent Structures in Ultrasound Images Degraded by Acoustic Clutter
No full textThis work demonstrates that the combination of multi-line transmission (MLT) and short-lag spatial coherence (SLSC) imaging improves the contrast of highly coherent structures within soft tissues when compared to both traditional SLSC imaging and conventional delay and sum (DAS) beamforming. Experimental tests with small (i.e., -3 mm) targets embedded in homogeneous and heterogeneous backgrounds were conducted. DAS or SLSC images were reconstructed when implementing MLT with varying numbers of simultaneously transmitted beams. In images degraded by acoustic clutter, MLT SLSC achieved up to 34.1 dB better target contrast and up to 16 times higher frame rates when compared to the more conventional single-line transmission SLSC images, with lateral resolution improvements as large as 38.2%. MLT SLSC thus represents a promising technique for clinical applications in which ultrasound visualization of highly coherent targets is required (e.g., breast microcalcifications, kidney stones, and percutaneous biopsy needle tracking) and would otherwise be challenging due to the strong presence of acoustic clutter
Improving the Quality of Monostatic Synthetic-Aperture Ultrasound Imaging Through Deep-Learning-Based Beamforming
No full textIn synthetic aperture (SA) ultrasound imaging, either monostatic or multistatic approaches can be employed. In both cases, in transmission, a single element of the transducer array is used at each time. In reception, the same element is used for the monostatic approach, while the whole array is used for the multistatic one. Thus, the monostatic approach could be implemented using a simpler single-channel architecture, however at the expense of image quality, while the multistatic one provides a high quality image but requires a more complex N-channel system. In this work, we show that a deep neural network can be trained to reconstruct images with a high contrast, as in the multistatic SA case (considering a 128-element array), but starting from the pre-beamforming signals acquired through the monostatic SA approach. We implemented a U-net and trained it using 27200 simulated signal-sets and the corresponding target images generated with Field II, considering numerical phantoms with random elliptical targets. The deep neural network (DNN) output image quality was evaluated in terms of contrast on a test set made of 500 simulated images, and on experimental scans of a commercial phantom and of the carotid artery. The results show that, after training over 39 epochs, the DNN is able to provide images with a good quality starting from the radiofrequency signals obtained with a simple monostatic SA approach, potentially requiring a single-channel only
Seduta regolabile per carrozzine destinate a soggetti diversamente abili
No full textIl brevetto riguarda un sistema meccanico per la regolazione delle zone di appoggio costituenti la seduta di una carrozzina per disabili. Il sistema consente l’adeguamento flessibile della seduta alle condizioni deficitarie dell’apparato muscolo-scheletrico del disabile
Deep Semantic Segmentation of Echocardiographic Images using Vision Transformers
No full textDeep learning (DL) methods have revolutionized image segmentation by providing tools to automatically identify structures within images, with high levels of accuracy. In particular, Convolutional Neural Networks (CNN), such as the U-Net and its variants, have achieved remarkable results in many segmentation tasks. Only recently, Vision Transformer (ViT)-based models have emerged and in some cases demonstrated to outperform CNNs in semantic segmentation tasks. However, transformers typically require larger amounts of data for training as compared to CNNs. This can result in a significant drawback given the time-consuming nature of collecting and annotating data, especially in a clinical setting. In particular, only a few studies involved the application of ViT networks for ultrasound image segmentation.In this study, we propose one of the earliest applications of ViT-based architectures for segmenting the left heart chambers in 2D echocardiographic images. Indeed, the identification of cardiac structures, like e.g. heart chambers, can be used to derive relevant quantitative parameters, such as atrial and ventricular volumes, the ejection fraction, etc.We trained and tested several ViT models using the publicly available CAMUS dataset, composed by cardiac image sequences from 500 patients, with the corresponding mask labels of the left ventricle and left atrium. ViT networks performances were then compared to an implementation of the U-Net. We demonstrate how, on this type of data, recent ViT variants can reach and even outperform CNNs, despite the limited data availability
- …
