<span id="docs-internal-guid-cdb76bbb-7fff-978d-961c-e21c41807064"><span>During the last few years, deep learning achieved remarkable results in the field of machine learning when used for computer vision tasks. Among many of its architectures, deep neural network-based architecture known as convolutional neural networks are recently used widely for image detection and classification. Although it is a great tool for computer vision tasks, it demands a large amount of training data to yield high performance. In this paper, the data augmentation method is proposed to overcome the challenges faced due to a lack of insufficient training data. To analyze the effect of data augmentation, the proposed method uses two convolutional neural network architectures. To minimize the training time without compromising accuracy, models are built by fine-tuning pre-trained networks VGG16 and ResNet50. To evaluate the performance of the models, loss functions and accuracies are used. Proposed models are constructed using Keras deep learning framework and models are trained on a custom dataset created from Kaggle CAT vs DOG database. Experimental results showed that both the models achieved better test accuracy when data augmentation is employed, and model constructed using ResNet50 outperformed VGG16 based model with a test accuracy of 90% with data augmentation &amp; 82% without data augmentation.</span></span>

Abstmct-This paper is a survey of recent literature in adaptive flight control. Although there are many published articles using classical control techniques, the trend appears to be toward using intelligent control techniques, particularly neural networks based methods. The survey is not presented in a chronological order, and it is not claimed to be exhaustive.In this rapidly progressing field, the only claim to be made is that it is a work in progress.

Abstmct-This paper is a survey of recent literature in neural networks applications in the field of automatic control. It is now generally accepted in the field that for nonlinear, imperfectly or partially known, and complicated systems, neural networks offer some of the most effective control techniques. In this paper, no attempt has been made to provide mathematical or algorithmic details of the various approaches that are being proposed in the literature; instead, general outlines of some of the techniques have been given. The survey is not presented in a chronological order. It is not an exhaustive survey -but an effort has been made to collect publications from different types of journals. Many authors and groups of authors in the field have numerous publications; for each group, we have cited only one or two representative articles. The goal of this paper is to serve as a resource for new researchers in the field.

No abstract

COVID-19 pandemic has posed serious risk of contagion to humans. There is a need to find reliable non-contact tests like vocal correlates of COVID-19 infection. Thirty-six Asian ethnic volunteers 16 (8M & 8F) infected subjects and 20 (10M &10F) non-infected controls participated in this study by vocalizing vowels /a/, /e/, /i/, /o/, /u/. Voice correlates of 16 COVID-19 positive patients were compared during infection and after recovery with 20 non-infected controls. Compared to non-infected controls, significantly higher values of energy intensity for /o/ ( p = 0.048); formant F1 for /o/ ( p = 0.014); and formant F3 for /u/ ( p = 0.032) were observed in male patients, while higher values of Jitter (local, abs) for /o/ ( p = 0.021) and Jitter (ppq5) for /a/ ( p = 0.014) were observed in female patients. However, formant F2 for /u/ ( p = 0.018), mean pitch F0 for /e/, /i/ and /o/ ( p = 0.033; 0.036; 0.047) decreased for female patients under infection. Compared to recovered conditions, HNR for /e/ ( p = 0.014) was higher in male patients under infection, while Jitter (rap) for /a/ ( p = 0.041); Jitter (ppq5) for /a/ ( p = 0.032); Shimmer (local, dB) for /i/ ( p = 0.024); Shimmer (apq5) for /u/ ( p = 0.019); and formant F4 for vowel /o/ ( p = 0.022) were higher in female patients under infection. However, HNR for /e/ ( p = 0.041); and formant F1 for /o/ ( p = 0.002) were lower in female patients compared to their recovered conditions. Obtained results support the hypothesis since changes in voice parameters were observed in the infected patients which can be correlated to a combination of acoustic measures like fundamental frequency, formant characteristics, HNR, and voice perturbations like jitter and shimmer for different vowels. Thus, voice analysis can be used for scanning and prognosis of COVID-19 infection. Based on the findings of this study, a mobile application can be developed to analyze human voice in real-time to detect COVID-19 symptoms for remedial measures and necessary action.