Learning acoustic models directly from the raw waveform data with minimal processing is challenging. Current waveform-based models have generally used very few (~2) convolutional layers, which might ...be insufficient for building high-level discriminative features. In this work, we propose very deep convolutional neural networks (CNNs) that directly use time-domain waveforms as inputs. Our CNNs, with up to 34 weight layers, are efficient to optimize over very long sequences (e.g., vector of size 32000), necessary for processing acoustic waveforms. This is achieved through batch normalization, residual learning, and a careful design of down-sampling in the initial layers. Our networks are fully convolutional, without the use of fully connected layers and dropout, to maximize representation learning. We use a large receptive field in the first convolutional layer to mimic bandpass filters, but very small receptive fields subsequently to control the model capacity. We demonstrate the performance gains with the deeper models. Our evaluation shows that the CNN with 18 weight layers outperforms the CNN with 3 weight layers by over 15% in absolute accuracy for an environmental sound recognition task and is competitive with the performance of models using log-mel features.
Window operation is not only an important method for improving the indoor thermal environment and air quality, but also a significant way to reduce energy consumption of air-conditioned rooms during ...off-running periods in transition seasons. The occupants' window-operation behavior is influenced by both objective factors, such as thermal comfort and indoor air quality; and objective sensation, such as psychology and physiology, introducing considerable randomness and uncertainty. A two-month field observation of occupant window-opening behaviors for natural ventilation in an office building during the transition seasons was carried out in Chongqing, China. Multi-factor analysis of variance was conducted in data analysis using SPSS statistical software. The results showed that outdoor air temperature significantly affected window opening among other factors such as outdoor relative humidity, indoor air temperature, indoor relative humidity, and indoor CO2 concentration, which have much less effect. The main trigger point for opening windows in the transition seasons is from occupants' desire to improve the indoor thermal and air quality environment. A probability model of occupants' window operation was proposed based on logistic regression analysis. Meanwhile, the Monte Carlo simulation results indicate that during transition seasons (when outdoor temperature varied from 15 to 30 °C), the probability of window opening in office buildings follows a normal distribution and increases linearly along with the outdoor temperature growth.
•We investigated window-opening behavior during the transition seasons when air-conditioner is inoperative.•Monte Carlo simulation method was used to predict the window-opening probability.•The probability of window-opening in the office building during transition seasons follows a normal distribution.•When the outdoor temperature is beyond the range of between 15 °C and 30 °C, bias exists.
Environmental sound detection is a challenging application of machine learning because of the noisy nature of the signal, and the small amount of (labeled) data that is typically available. This work ...thus presents a comparison of several state-of-the-art Deep Learning models on the IEEE challenge on Detection and Classification of Acoustic Scenes and Events (DCASE) 2016 challenge task and data, classifying sounds into one of fifteen common indoor and outdoor acoustic scenes, such as bus, cafe, car, city center, forest path, library, train, etc. In total, 13 hours of stereo audio recordings are available, making this one of the largest datasets available.
Single image dehazing is a challenging ill-posed problem that has drawn significant attention in the last few years. Recently, convolutional neural networks have achieved great success in image ...dehazing. However, it is still difficult for these increasingly complex models to recover accurate details from the hazy image. In this paper, we pay attention to the feature extraction and utilization of the input image itself. To achieve this, we propose a Multi-scale Topological Network (MSTN) to fully explore the features at different scales. Meanwhile, we design a Multi-scale Feature Fusion Module (MFFM) and an Adaptive Feature Selection Module (AFSM) to achieve the selection and fusion of features at different scales, so as to achieve progressive image dehazing. This topological network provides a large number of search paths that enable the network to extract abundant image features as well as strong fault tolerance and robustness. In addition, ASFM and MFFM can adaptively select important features and ignore interference information when fusing different scale representations. Extensive experiments are conducted to demonstrate the superiority of our method compared with state-of-the-art methods.
The construction of trigonometric interpolatory splines plays a very important role in geometric modeling. This paper presents a quartic trigonometric interpolatory spline with local free parameters. ...The new spline not only automatically interpolates the given points and achieves C2 continuity, but also owns shape adjustability when the points remain fixed. Some examples show that the shape of the new spline can easily realize local and global adjustment by changing the free parameters.
A new approach to enriching britholite phase from the rare-earth-rich slag by super gravity was investigated. The Bayan Obo iron ore, which was used as raw material, was reduced and melting separated ...to produce iron nugget and rare-earth-rich slag. Subsequently, the slag was heat-treated and enriched in the super gravity filed. The volume fraction and equivalent diameter of britholite phase were measured by scanning electron microscope (SEM) and image analyzer, whereas the mineral composition and chemical component were characterized by X-ray diffraction and X-ray fluorescence. The results indicated that the samples obtained by the gravity coefficient
G
≥ 500,
t
≥ 15 minutes, and
T
≥1423 K (1150 °C) show significant layers and britholite phase present gradient size distribution in the sample along the super gravity. The layered sample was central cut and characterized by SEM, and it is difficult to find any britholite particles in the upper area of the sample. The britholite phase gathers at the middle and bottom areas of the sample. The mechanism of moving speed of britholite particles in super gravity field was discussed, and the conclusion indicates that the moving speed of britholite particles is proportional to the square of the britholite particle size. As a result, large britholite particles move farther than the small ones and gather at the bottom of the sample, whereas small britholite particles accumulate in the middle of the sample. Under the hypothesis that rare earth (RE) exists in the slag in terms of RE
2
O
3,
with the gravity coefficient
G
= 500,
t
= 15 minutes, and
T
= 1423 K (1150 °C), the mass fraction of RE
2
O
3
in the concentrate is up to 23.29 pct whereas that of the tailing is just 5.57 pct. Considering that the mass fraction of RE
2
O
3
is 12.01 pct in the parallel sample, the recovery ratio of RE in the concentrate is up to 71.19 pct by centrifugal enrichment.
Cubic Hermite interpolation curve plays a very important role in interpolation curves modeling, but it has three shortcomings including low continuity, difficult shape adjustment, and the inability ...to accurately represent some common engineering curves. We construct a cubic trigonometric Hermite interpolation curve to make up the three shortcomings of cubic Hermite interpolation curve once and for all. The cubic trigonometric Hermite interpolation curve not only inherits the features of cubic Hermite interpolation curve but also achieves C2 continuity, has local and global adjustability, and can accurately represent elliptical arc, circular arc, quadratic parabolic arc, cubic parabolic arc, and astroid arc that often appear in engineering. In addition, we give the schemes for optimizing the shape of the cubic trigonometric Hermite interpolation curve based on internal energy minimization. The schemes include optimizing the shape of planar curve and spatial curve. Some modeling examples show that the proposed schemes are effective and the cubic trigonometric Hermite interpolation curve is more practical than cubic Hermite interpolation curve.
In this paper, we aim at smoothing two connected ball Bézier curves from Cr−1 to Crr≥1 by minimizing the energies of the curves. We propose the algorithms based on internal energy minimization and ...curve attractor minimization. Then, we combine the internal energy and the curve attractor and give the algorithm based on combined energy minimization. All algorithms are established by solving bi-objective minimizations. Some numerical examples show that the proposed algorithms are effective, making them useful for smoothing 3D objects constructed by connected ball Bézier curves.
In this letter, we present a novel approach to efficiently generate collision-free optimal trajectories for multiple non-holonomic mobile robots in obstacle-rich environments. Our approach first ...employs a graph-based multi-agent path planner to find an initial discrete solution, and then refines this solution into smooth trajectories using nonlinear optimization. We divide the robot team into small groups and propose a prioritized trajectory optimization method to improve the scalability of the algorithm. Infeasible sub-problems may arise in some scenarios because of the decoupled optimization framework. To handle this problem, a novel grouping and priority assignment strategy is developed to increase the probability of finding feasible trajectories. Compared to the coupled trajectory optimization, the proposed approach reduces the computation time considerably with a small impact on the optimality of the plans. Simulations and hardware experiments verified the effectiveness and superiority of the proposed approach.
Papillary thyroid carcinoma (PTC) is the most common pathological type of thyroid malignancy and also has an excellent prognosis. Primary thyroid lymphoma (PTL) is rare and has a poor prognosis. The ...co-occurrence of both malignancies is extremely rare, and the preoperative diagnosis is rather difficult. We report the case of a patient with both PTC and PTL in the setting of Hashimoto’s thyroiditis (HT). A 59-year-old female patient was referred to our department for progressive enlargement of the thyroid gland over a few months. The imaging results demonstrated an enlarged thyroid and a mass in the thyroid. Total thyroidectomy and bilateral central neck node dissection were conducted. The final diagnosis of the coexistence of thyroid diffuse large B cell lymphoma and PTC was confirmed by histopathology and immunohistochemistry. The patient received radiation therapy and six cycles of chemotherapy combined with targeted therapy, including rituximab, cyclophosphamide, doxorubicin, vindesine, and prednisone (R-CHOP). After 6 months of follow-up, neither tumor has recurred. It is important for physicians to keep PTL in mind for differential diagnosis in HT patients with sudden thyroid enlargement.