•Testing several Convolutional Neural Networks on for skeletal bone age assessment with X-Ray images.•BoNet: a CNN for automated skeletal age assessment able to cope with hand nonrigid ...deformation.•First automated skeletal bone age assessment work tested on a public dataset with source code publicly available.•Providing answers to more general questions about deep learning on medical images.
Display omitted
Skeletal bone age assessment is a common clinical practice to investigate endocrinology, genetic and growth disorders in children. It is generally performed by radiological examination of the left hand by using either the Greulich and Pyle (G&P) method or the Tanner–Whitehouse (TW) one. However, both clinical procedures show several limitations, from the examination effort of radiologists to (most importantly) significant intra- and inter-operator variability. To address these problems, several automated approaches (especially relying on the TW method) have been proposed; nevertheless, none of them has been proved able to generalize to different races, age ranges and genders.
In this paper, we propose and test several deep learning approaches to assess skeletal bone age automatically; the results showed an average discrepancy between manual and automatic evaluation of about 0.8 years, which is state-of-the-art performance. Furthermore, this is the first automated skeletal bone age assessment work tested on a public dataset and for all age ranges, races and genders, for which the source code is available, thus representing an exhaustive baseline for future research in the field.
Beside the specific application scenario, this paper aims at providing answers to more general questions about deep learning on medical images: from the comparison between deep-learned features and manually-crafted ones, to the usage of deep-learning methods trained on general imagery for medical problems, to how to train a CNN with few images.
In this work, we propose a 3D fully convolutional architecture for video saliency prediction that employs hierarchical supervision on intermediate maps (referred to as
conspicuity maps
) generated ...using features extracted at different abstraction levels. We provide the base hierarchical learning mechanism with two techniques for
domain adaptation
and
domain-specific learning
. For the former, we encourage the model to unsupervisedly learn hierarchical general features using gradient reversal at multiple scales, to enhance generalization capabilities on datasets for which no annotations are provided during training. As for domain specialization, we employ domain-specific operations (namely, priors, smoothing and batch normalization) by specializing the learned features on individual datasets in order to maximize performance. The results of our experiments show that the proposed model yields state-of-the-art accuracy on supervised saliency prediction. When the base hierarchical model is empowered with domain-specific modules, performance improves, outperforming state-of-the-art models on three out of five metrics on the DHF1K benchmark and reaching the second-best results on the other two. When, instead, we test it in an unsupervised domain adaptation setting, by enabling hierarchical gradient reversal layers, we obtain performance comparable to supervised state-of-the-art. Source code, trained models and example outputs are publicly available at
https://github.com/perceivelab/hd2s
.
Microstructural aspects in Al–Cu dissimilar joining by FSW Carlone, Pierpaolo; Astarita, Antonello; Palazzo, Gaetano S. ...
International journal of advanced manufacturing technology,
07/2015, Letnik:
79, Številka:
5-8
Journal Article
Recenzirano
Sound AA2024-T3–Cu10100 dissimilar joints were obtained by friction stir welding offsetting the tool probe towards the aluminum sheet and employing selected processing parameters. Joint ...microstructure was analyzed by means of conventional optic microscopy as well as scanning electron microscopy. The weld bead exhibited welding zones and some features typically encountered in similar FSW. The nugget zone consisted of a mixture of recrystallized aluminum matrix and deformed and twinned copper particles. Onion rings and particle-rich zones, made of Cu particles dispersed in the Al matrix, were also observed. EDS analysis revealed that several Al–Cu intermetallic compounds, such as Al
2
Cu, AlCu, and Al
3
Cu
4
, chemically different w.r.t. compounds precipitated during the T3 aging treatment (Al
3
Cu), were formed during the process. Microstructure variation significantly affects the microhardness distribution in the cross-section of the joint.
Human behavior understanding in videos is a complex, still unsolved problem and requires to accurately model motion at both the local (pixel-wise dense prediction) and global (aggregation of motion ...cues) levels. Current approaches based on supervised learning require large amounts of annotated data, whose scarce availability is one of the main limiting factors to the development of general solutions. Unsupervised learning can instead leverage the vast amount of videos available on the web and it is a promising solution for overcoming the existing limitations. In this paper, we propose an adversarial GAN-based framework that learns video representations and dynamics through a self-supervision mechanism in order to perform dense and global prediction in videos. Our approach synthesizes videos by (1) factorizing the process into the generation of static visual content and motion, (2) learning a suitable representation of a motion latent space in order to enforce spatio-temporal coherency of object trajectories, and (3) incorporating motion estimation and pixel-wise dense prediction into the training procedure. Self-supervision is enforced by using motion masks produced by the generator, as a co-product of its generation process, to supervise the discriminator network in performing dense prediction. Performance evaluation, carried out on standard benchmarks, shows that our approach is able to learn, in an unsupervised way, both local and global video dynamics. The learned representations, then, support the training of video object segmentation methods with sensibly less (about 50%) annotations, giving performance comparable to the state of the art. Furthermore, the proposed method achieves promising performance in generating realistic videos, outperforming state-of-the-art approaches especially on motion-related metrics.
•Background and foreground modeling method able to run seamlessly under extreme conditions.•Modeling structural variations of pixels’ neighbors via joint domain-range approach integrating textons ...into the model.•Exhaustive testing of state-of-the-art approaches on real-life scenarios.
Background modeling is a well-know approach to detect moving objects in video sequences. In recent years, background modeling methods that adopt spatial and texture information have been developed for dealing with complex scenarios. However, none of the investigated approaches have been tested under extreme conditions, such as the underwater domain, on which effects compromising the video quality affect negatively the performance of the background modeling process. In order to overcome such difficulties, more significant features and more robust methods must be found. In this paper, we present a kernel density estimation method which models background and foreground by exploiting textons to describe textures within small and low contrasted regions. Comparison with other texture descriptors, namely, local binary pattern (LBP) and scale invariant local ternary pattern (SILTP) shown improved performance. Besides, quantitative and qualitative performance evaluation carried out on three standard datasets showing very complex conditions revealed that our method outperformed state-of-the-art methods that use different features and modeling techniques and, most importantly, it is able to generalize over different scenarios and targets.
xWhat if we could effectively read the mind and transfer human visual capabilities to computer vision methods? In this paper, we aim at addressing this question by developing the first visual object ...classifier driven by human brain signals. In particular, we employ EEG data evoked by visual object stimuli combined with Recurrent Neural Networks (RNN) to learn a discriminative brain activity manifold of visual categories in a reading the mind effort. Afterward, we transfer the learned capabilities to machines by training a Convolutional Neural Network (CNN)-based regressor to project images onto the learned manifold, thus allowing machines to employ human brain-based features for automated visual classification. We use a 128-channel EEG with active electrodes to record brain activity of several subjects while looking at images of 40 ImageNet object classes. The proposed RNN-based approach for discriminating object classes using brain signals reaches an average accuracy of about 83%, which greatly outperforms existing methods attempting to learn EEG visual object representations. As for automated object categorization, our human brain-driven approach obtains competitive performance, comparable to those achieved by powerful CNN models and it is also able to generalize over different visual datasets.
Liquid composite molding processes are manufacturing techniques involving the impregnation and saturation of dry fibrous preforms by means of injection or infusion of catalyzed resin systems. ...Complete wetting of the reinforcement and reduction of voids are key issues to enhance mechanical properties of the final product, as a consequence on line monitoring and control of resin flow is highly desirable to detect and avoid potentialbet macro- as well as micro-voids. In this paper, parallel-plate dielectric sensors were investigated to track the position of unsaturated as well as saturated flow fronts through dual scale porous media. Sensors configuration was analyzed and improved via electromagnetic (EM) finite element simulations. The effectiveness of the proposed system was assessed in one-dimensional impregnation tests. Good agreement was found between unsaturated front positions provided by the considered system and acquired through conventional visual techniques. An indirect verification strategy, based on CFD and EM simulations of the process, was applied to investigate the reliability of dielectric sensors with respect to saturation phenomena. Obtained outcomes highlighted the intriguing capabilities of the proposed method.
The quantification of stenosis severity from X-ray catheter angiography is a challenging task. Indeed, this requires to fully understand the lesion's geometry by analyzing dynamics of the contrast ...material, only relying on visual observation by clinicians. To support decision making for cardiac intervention, we propose a hybrid CNN-Transformer model for the assessment of angiography-based non-invasive fractional flow-reserve (FFR) and instantaneous wave-free ratio (iFR) of intermediate coronary stenosis. Our approach predicts whether a coronary artery stenosis is hemodynamically significant and provides direct FFR and iFR estimates. This is achieved through a combination of regression and classification branches that forces the model to focus on the cut-off region of FFR (around 0.8 FFR value), which is highly critical for decision-making. We also propose a spatio-temporal factorization mechanisms that redesigns the transformer's self-attention mechanism to capture both local spatial and temporal interactions between vessel geometry, blood flow dynamics, and lesion morphology. The proposed method achieves state-of-the-art performance on a dataset of 778 exams from 389 patients. Unlike existing methods, our approach employs a single angiography view and does not require knowledge of the key frame; supervision at training time is provided by a classification loss (based on a threshold of the FFR/iFR values) and a regression loss for direct estimation. Finally, the analysis of model interpretability and calibration shows that, in spite of the complexity of angiographic imaging data, our method can robustly identify the location of the stenosis and correlate prediction uncertainty to the provided output scores.
Summary
Suctioning is essential in managing tracheal tubes, but also has drawbacks. Using a bench model, we demonstrated the extent and time course of pressure changes during suctioning, examined ...their relationship with tracheal tube and catheter diameters and assessed the effects of artificial ‘sputum’ and of compensatory gas flow in the system. We suctioned at −20 kPa (−150 mmHg) and −80 kPa (−600 mmHg) using three different sized catheters and a 5.9‐mm diameter bronchoscope through tracheal tubes ranging from 6.5 mm to 9.0 mm in diameter. Pressure changes ranged from −0.1 kPa (−0.8 mmHg) to −20.4 kPa (−153.0 mmHg). We demonstrated more negative pressures with decreasing tracheal tube diameter (p = 0.024) and increasing catheter diameter (p = 0.038). Addition of artificial ‘sputum’ led to more negative, but unpredictable, pressures than those seen with clean tubes (p = 0.012). Bronchoscopic suctioning produced pressure changes even greater than the largest suction catheter (p = 0.0039). Using a closed system with continuous positive airway pressure and 155 l.min−1 compensatory gas flow attenuated the pressure changes generated both with a 4.0‐mm catheter (p = 0.0005) and on bronchoscopic suctioning (p = 0.0078). The time taken to reach 50% of minimum pressure was always less than 1 s. The use of high compensatory flows during suctioning merits clinical evaluation.
Previous analyses of the GIM (Gruppo Italiano Mammella) 2 study showed that addition of fluorouracil to epirubicin, cyclophosphamide, and paclitaxel in patients with node-positive early breast cancer ...does not improve outcome, whereas dose-dense chemotherapy induces a significant improvement in both disease-free survival and overall survival as compared with a standard schedule. Here, we present long-term results of the study.
In this 2 × 2 factorial, open-label, randomised, phase 3 trial, we enrolled patients aged 18–70 years with operable, node-positive, breast cancer with Eastern Cooperative Oncology Group performance status of 0–1 from 81 hospitals in Italy. Eligible patients were randomly allocated (1:1:1:1) to one of the four following study groups: four cycles of standard-interval intravenous EC (epirubicin 90 mg/m2 and cyclophosphamide 600 mg/m2) on day 1 every 3 weeks, followed by four cycles of intravenous paclitaxel (175 mg/m2) on day 1 every 3 weeks (q3EC-P group); four cycles of intravenous FEC (fluorouracil 600 mg/m2, epirubicin 90 mg/m2, and cyclophosphamide 600 mg/m2) on day 1 every 3 weeks, followed by four cycles of intravenous paclitaxel (175 mg/m2) on day 1 every 3 weeks (q3FEC-P group); dose-dense EC-P regimen, with the same doses and drugs as the q3EC-P group but administered every 2 weeks (q2EC-P group); and the dose-dense FEC-P regimen, with the same doses and drugs as the q3FEC-P group but given every 2 weeks (q2FEC-P). Randomisation, with stratification by centre, with permuted blocks of size 12, was done with a centralised, interactive, internet-based system that randomly generated the treatment allocation. The primary endpoint was disease-free survival in the intention-to-treat population, comparing different chemotherapy schedule (dose-dense vs standard-dose intervals) and regimen (FEC-P vs EC-P). Safety population included all patients that received at least one dose of any study drug according to the treatment received. This trial is registered with ClinicalTrials.gov, NCT00433420, and is now closed.
Between April 24, 2003, and July 3, 2006, 2091 patients were randomly assigned to treatment: 545 to q3EC-P, 544 to q3FEC-P, 502 to q2EC-P, and 500 to q2FEC-P. 88 patients were enrolled in centres providing only standard interval schedule and were assigned only to q3FEC-P and q3EC-P; thus, 2091 patients were included in the intention-to-treat analysis for the comparison of EC-P (1047 patients) versus FEC-P (1044 patients) and 2003 patients were included in the intention-to-treat analysis for the comparison of dose-dense (1002 patients) versus standard interval analysis (1001 patients). After a median follow-up of 15·1 years (IQR 8·4–16·3), median disease-free survival was not significantly different between FEC-P and EC-P groups (17·09 years 95% CI 15·51–not reached vs not reached 17·54–not reached; unadjusted hazard ratio 1·12 95% CI 0·98–1·29; log-rank p=0·11). Median disease-free survival was significantly higher in the dose-dense interval group than the standard-interval group (not reached 95% CI 17·45–not reached vs 16·52 14·24–17·54; 0·77 95% CI 0·67–0·89; p=0·0004). The most common grade 3–4 adverse events were neutropenia (200 37% of 536 patients in the q3EC-P group vs 257 48% of 533 in the q3FEC-P group vs 50 10% of 496 q2EC-P vs 97 20% of 492) and alopecia (238 44% vs 249 47% vs 228 46% vs 235 48%). During extended follow-up, no further grade 3–4 adverse events or deaths related to toxic-effects were reported. Treatment-related serious adverse events were reported in nine (2%) patients in the q3EC-P group, seven (1%) in the q3FEC-P group, nine (2%) in the q2EC-P group, and nine (2%) in the q2FEC-P group. No treatment-related deaths occurred.
Updated results from the GIM2 study support that optimal adjuvant chemotherapy for patients with high-risk early breast cancer should not include fluorouracil and should use a dose-dense schedule.
Bristol-Myers Squibb, Pharmacia, Dompè Biotec Italy, Italian Ministry of Health, Fondazione Italiana per la Ricerca sul Cancro, and Alliance Against Cancer.