An efficient visual aid is essential for visually impaired candidate. At the same instance it has to be simple, robust and cost effective. However, regardless of being expensive, it is challenging to incorporate high ...
An efficient visual aid is essential for visually impaired candidate. At the same instance it has to be simple, robust and cost effective. However, regardless of being expensive, it is challenging to incorporate high end artificial intelligence into a compact device to serve the purpose. This paper presents a prototype of low-cost smart glasses for visually impaired individuals. Also, a brief survey on the cost and features of the goggles available in the market are included. In this work a prototype is developed integrating Raspberry Pi module, camera, sensors and a goggles. The captured images are processed using object recognition models and its corresponding text is read aloud using google text to translator. Subsequently, a comparative analysis of object recognition model is explored and discussed. Essential experimental results are incorporated along with cost involved for the prototype development.
In this paper, a method to detect arbitrary driving scenes from actual driving behavior data is proposed. This is positioned as a necessary technology to evaluate a driver’s cognitive ability from actual driving beha...
In this paper, a method to detect arbitrary driving scenes from actual driving behavior data is proposed. This is positioned as a necessary technology to evaluate a driver’s cognitive ability from actual driving behavior data. In order to evaluate a driver’s cognitive ability from actual driving behavior data, it is first necessary to be able to detect scenes in which cognitive ability can be evaluated. However, since there is no clear definition of the scenes in which cognitive ability can be evaluated. Therefore, this method detects arbitrary driving scenes defined by the *** method is a binary classifier generation method based on deep learning. However, a driving scene is considered to be composed of road information, appearing object information, and operation information, and each is expressed by four types of modalities: driving video, object label vector, grid feature, and operation data in this method. By adopting a deep learning model structure that incorporates the idea of stacking ensemble learning and training a binary classifier based on four types of modal information, a detector that detects the desired driving situation is generated. In this paper, the effectiveness of the proposed method through experiments is verified.
This paper focuses on solving a stochastic saddle point problem (SPP) under an overparameterized regime for the case, when the gradient computation is impractical. As an intermediate step, we generalize Same-sample St...
详细信息
The matrix factorization algorithm has achieved good results in the field of scenic spot recommendation, but there are still many problems. Most of the studies only refer to the user’ s score of the tourist attractio...
详细信息
ISBN:
(纸本)9798350309720
The matrix factorization algorithm has achieved good results in the field of scenic spot recommendation, but there are still many problems. Most of the studies only refer to the user’ s score of the tourist attraction to represent the user ’s overall evaluation of the attraction, thus ignoring the correlation between the online review text information of the attraction and the attraction, which eventually leads to the low accuracy of the entire recommendation system. This paper proposes a personalized tourism recommendation algorithm that integrates scenic spot labels and sentiment polarity analysis. On the basis of constructing feature labels for various scenic spots, the sentiment polarity analysis of user comments is carried out through natural language text information processing, and implicit emotions are given to the labels contained in scenic spots. Secondly, it is introduced into the factor vector of matrix decomposition model algorithm, and finally the potential relationship between users and feature labels and scenic spots is deeply explored. Through the user ’s preference for emotional feature labels, the user ’s approximate rating value for attractions is predicted. Experiments show that compared with the user-based collaborative filtering scenic spot recommendation algorithm, the average absolute error (MAE) and root mean square error (RMSE) are reduced by 70.28 % and 65.23 % respectively. Compared with the tag-based collaborative filtering attraction recommendation algorithm, MAE and RMSE are reduced by 65.02 % and 63.93 %, respectively. Compared with the matrix factorization recommendation algorithm of scenic spot labels, MAE and RMSE are reduced by 34.02% and 29.93 % respectively. To sum up the performance of the algorithm is significantly higher than the existing attractions recommended algorithm can be more accurate to provide users with recommended options.
In recent years, the application of radiofrequency identification technology in smart access control systems has become a trend. Users only need to take out the RFID electronic key and approach the sensor of the gate ...
In recent years, the application of radiofrequency identification technology in smart access control systems has become a trend. Users only need to take out the RFID electronic key and approach the sensor of the gate to open the gate through signal induction. For housewives who always have a lot of bags and bags in their hands, there is no need to free up one hand to find the key to open the door and avoid being in a hurry. However, it is relatively inconvenient for drivers. Instead, it is much more convenient to use wireless controllers or license plate recognition. This study proposes a system that combines RFID with optical communication technology and vehicle headlights to open the garage door with a light code.
In recent times, the number of laryngeal cancer patients has increased significantly around the world. Proper care for laryngeal cancer is complex, particularly in the last phase. This cancer category is a complicated...
详细信息
In recent times, the number of laryngeal cancer patients has increased significantly around the world. Proper care for laryngeal cancer is complex, particularly in the last phase. This cancer category is a complicated tumour in the head and neck regions of the individual. Recently, investigators have established different diagnosis methods and devices to assist medical experts in recognizing laryngeal cancer efficiently. Endoscopic schemes with narrow-band imaging (NBI), which improves the visualization of subepithelial and epithelial microvascular patterns, attain a high specificity or sensitivity in the initial recognition of laryngeal cancer. The computer-aided system uses artificial intelligence (AI) over machine learning (ML) and deep learning (DL), containing a convolutional neural network (CNN), for automatic disease detection and diagnosis. Therefore, this study presents a Fusion of Efficient Transfer Learning Models with Pelican Optimization for Accurate Laryngeal Cancer Detection and Classification (FETLM-POALCDC) method. The main intention of the FETLM-POALCDC method is to facilitate the classification and recognition of laryngeal cancer in throat region images. Initially, the FETLM-POALCDC method employs the Wiener filter (WF) model in the image preprocessing to eliminate noise while preserving critical image details. Furthermore, the fusion of transfer learning comprising three models, EfficientNetV2, NASNetMobile, and ResNet152, is employed to capture multiscale and discriminative features. Integrating convolutional neural networks with bidirectional gated recurrent units (CNN+BiGRU) classifier is utilized to detect laryngeal cancer. To enhance the hyperparameters of the CNN+BiGRU method, the pelican optimization algorithm (POA) is employed to ensure improved performance by fine-tuning essential parameters. A series of empirical studies of the FETLM-POALCDC approach is performed under the Laryngeal dataset. The experimental validation of the FETLM-POAL
Developed since 1968, Puskesmas is the most important healthcare facility and is at the forefront of providing basic healthcare services at the community level. Basically, its role is very important and should be one ...
详细信息
The two-dimensional electron gas at the interface between LaAlO3 and SrTiO3 (LAO/STO) exhibits gate-tunable superconductivity with a characteristic domelike shape of the critical temperature (Tc) in the phase diagram....
详细信息
The two-dimensional electron gas at the interface between LaAlO3 and SrTiO3 (LAO/STO) exhibits gate-tunable superconductivity with a characteristic domelike shape of the critical temperature (Tc) in the phase diagram. As shown recently [Phys. Rev. B 102, 085420 (2020)], such an effect can be explained as a consequence of the extended s−wave symmetry of the gap within an intersite real space pairing scenario, leading to a good agreement between the experiment and theory. In this work, we turn to a detailed analysis of the influence of spin-orbit coupling on the LAO/STO phase diagram by considering separately the atomic component as well as the interorbital hopping induced by the broken inversion symmetry at the interface. In particular, we analyze the optimal carrier concentration for which the maximal Tc is reached relative to the Lifshitz transition point. We find that the misalignment between the two can be significantly enhanced by the spin-orbit splitting of the bands, combined with the fact that superconductivity sets in when the Fermi level passes the anticrossing induced by the spin-orbital hybridization. In the presence of the external in-plane magnetic field, our calculations show fourfold anisotropy with the paramagnetic limit largely exceeded for B|| directed along the high symmetry points [01] and [01]. The obtained electron concentration dependence of Bc|| reproduces the characteristic domelike shape reported in experiments and the estimated value of Bc|| corresponds to that measured experimentally.
We investigate correlated three-electron escape in Ne when driven by an intense, infrared laser field. We do so by employing a reduced-dimensionality quantum-mechanical model and two three-dimensional semi-classical m...
详细信息
Measuring the electromagnetic overall performance and configuration of networks in high-interference environments can be challenging. In those environments, the indicators of hobby can be infected with unwanted herita...
详细信息
ISBN:
(数字)9798350370249
ISBN:
(纸本)9798350370270
Measuring the electromagnetic overall performance and configuration of networks in high-interference environments can be challenging. In those environments, the indicators of hobby can be infected with unwanted heritage indicators, i.e., noise, leading to degraded overall performance and misguided size readings. This paper investigates the most reliable techniques for measuring networks’ electromagnetic performance and configuration in these problematic environments, utilizing introducing a unique multistage framework that combines sign-processing strategies with some community parameters. In particular, the framework includes four degrees: sign reception, sensitivity analysis, pass-correlation, and community configuration optimization-an analysis of the results of varying environmental conditions on sign reception and sensitivity change into conduct. A cross-correlation approach has evolved to decide the extent of correlation among one-of-a-kind alerts and to locate the specific configurations of the particular networks. Sooner or later, an optimization set of rules was carried out to locate the network’s optimum configuration. The final result of this research is a reliable method for measuring the electromagnetic performance and configuration of networks in high-interference environments. The proposed technique is anticipated to help researchers and engineers better understand the electromagnetic behaviors of complicated networks in actual-world scenarios.
暂无评论