2,655 research outputs found
Erratum: Hybrid group recommendation using modified termite colony algorithm: A context towards big data (Journal of Physical Chemistry (2018) 17:2 (1850019) DOI: 10.1142/S0219649218500193)
No full textWe would like to make the following correction to this article. The third author a±liation should be read as follows: Chintan Bhatt U. & P. U. Patel Department of Computer Engineering Charotar University of Science and Technology Changa, Gujarat 388421, India [email protected]
Emotion Unleashed: Real-Time FER in Video via Advanced Deep Learning Models
No full textWhile conversing, showing emotions through facial expressions comes naturally to human beings. Face expression detection is tackled by computer vision to serve psychology, human-computer interaction, and security applications. In this work, we present the software we have designed and implemented to read the emotions on people’s faces. The system allows users to discern emotions on people’s faces using artificial intelligence. It is a challenging task that has piqued the interest of scholars worldwide in facial expression analysis in recent years. Regarding some technical aspects of the proposed method, we rely on Deep learning methods, which have proved robust in various scenarios, including facial emotion recognition. This paper introduces a Deep Learning-based approach that relies on a Convolutional Neural Network (CNN) to infer knowledge from facial images and distinguish features. These features are then used to assign a label to an emotion. Deep Learning algorithms, such as DeepFace, NASNet Mobile, EfficientNet V2, and Inception V2, are investigated on FER tasks. By analyzing data from the FER-13 and Face databases, our model has learned to recognize emotions
Recommended from our members
The IAAF’s hyperandrogenism regulations suspended
No full textOn 27 July, the Court of Arbitration for Sport (‘CAS’) delivered a landmark ruling on the regulation of gender in sport. The decision explores how the categorisation of sport on the basis of sex can be best reconciled with the “biological reality” that human sex cannot necessarily be divided so clearly. Dr. Seema Patel, Senior Lecturer at Nottingham Trent University, Deputy Director of the Centre for Sports Law and author of ‘Inclusion and Exclusion in Competitive Sport: Socio-Legal and Regulatory Perspectives,’ reviews the case and suggests that sport regulation must be cautious of traditional criteria to determine eligibility in sports
Author interview: Q and A with Dr Ian Sanjay Patel on we’re here because you were there: immigration and the end of empire
No full textIn this author interview, we speak to Dr Ian Sanjay Patel about his new book, We’re Here Because You Were There: Immigration and the End of Empire, which explores post-war immigration laws, the afterlives of British imperial citizenship and related attempts to reimagine and rejuvenate British imperialism after 1945. Contributing to transnational histories of decolonisation, the book also explores the interconnections between human rights, post-war migration and international diplomacy. Author Interview with Dr Ian Sanjay Patel, author of We’re Here Because You Were There: Immigration and the End of Empire. Verso. 2021
Thermosensitive nanohydrogel of 5-fluorouracil for head and neck cancer: preparation, characterization and cytotoxicity assay
No full textChintan Dalwadi, Gayatri Patel Department of Pharmaceutics and Pharmaceutical Technology, Ramanbhai Patel College of Pharmacy, Charotar University of Science and Technology, Changa, Gujarat, India Abstract: Systemic chemotherapy has been shown to produce side effects. A small fraction of the drug reaches the tumor site; other healthy organs or normal tissues get affected or damaged due to the nonspecific action of these cytotoxic agents. Furthermore, due to their short period of activity, repeat injections are often required, which can lead to the exacerbation of side effects and inconvenience. To overcome these obstacles, in this study, we developed controlled and targeted intratumoral injection. Hydrogel was prepared by physical cross-linking method; however, nanohydrogel was prepared using tip probe-sonicator method. Our results revealed that biodegradable and thermosensitive 5-fluorouracil-loaded methylcellulose nanohydrogel synthesized by physical cross-linking method may be a beneficial approach in targeting the therapeutic agent to the tumor site. Keywords: biodegradable, intra tumor, targeted drug deliver
Embedded in the Body: the Poetry, History and Politics of Migritude with Shailja Patel (2021-02-25)
No full textOnline discussion, reading and Q&A; Thursday, February 25 at 4:00PM CST; Shailja Patel is the bestselling author of Migritude, taught in over 100 colleges and universities worldwide. Patel's poems have been translated into 17 languages, and been featured in the Smithsonian. The Nobel Women's Initiative honored her with a Global Feminist Spotlight. She is currently a Research Associate at Five College Women's Studies Research Center.Women, Gender & Sexuality Studies program; Alworth Institute for International Studies; Department of Anthropology, Sociology & Criminology; English program; Writing Studies programPatel, Shailja. (2021). Embedded in the Body: the Poetry, History and Politics of Migritude with Shailja Patel (2021-02-25). Retrieved from the University Digital Conservancy, https://hdl.handle.net/11299/220654
Lung sound disease detection using attention over pre-trained efficientnet architecture
No full textAir pollution has been on the rise for quite a while now and with it is the increasing number of cases involving respiratory diseases. These respiratory diseases range from the mild ones to the most severe ones. Therefore, we need to detect these diseases. The most efficient way of checking the health of the lung is through the use of lung sounds. But the lung sounds carry with them the minute variations that a doctor or any human sometimes misses. To tackle the same, this paper proposes a deep learning model that detects a person’s lung disease by processing lung sounds. The ICBHI 2017 dataset comprising the lung sound audios and their corresponding labels is used to train the proposed model. The dataset consists of lung sounds and some additional background noise has been removed using the bandpass Butterworth filter, whose frequency range has been set to 250Hz - 2000Hz. Later, the amount of data is multiplied through data augmentation for audio signals. The augmented data is later converted to mel spectrograms, which are, in turn, fed into the deep learning model. The model that gives the best accuracy is EfficientNet-B0 + attention with a training accuracy of 99.72%, validation accuracy of 99.82%, precision, recall and f1-scores of 99.82% as well. Furthermore, a comparison of the pre-trained model’s training and inference times has been performed. Although the proposed model takes a significant amount of training time, it takes the least amount of inference time compared to the other pre-trained models that have been tested
Enhancing Fingerprint Liveness Detection Accuracy Using Deep Learning: A Comprehensive Study and Novel Approach
No full textLiveness detection for fingerprint impressions plays a role in the meaningful prevention of any unauthorized activity or phishing attempt. The accessibility of unique individual identification has increased the popularity of biometrics. Deep learning with computer vision has proven remarkable results in image classification, detection, and many others. The proposed methodology relies on an attention model and ResNet convolutions. Spatial attention (SA) and channel attention (CA) models were used sequentially to enhance feature learning. A three-fold sequential attention model is used along with five convolution learning layers. The method’s performances have been tested across different pooling strategies, such as Max, Average, and Stochastic, over the LivDet-2021 dataset. Comparisons against different state-of-the-art variants of Convolutional Neural Networks, such as DenseNet121, VGG19, InceptionV3, and conventional ResNet50, have been carried out. In particular, tests have been aimed at assessing ResNet34 and ResNet50 models on feature extraction by further enhancing the sequential attention model. A Multilayer Perceptron (MLP) classifier used alongside a fully connected layer returns the ultimate prediction of the entire stack. Finally, the proposed method is also evaluated on feature extraction with and without attention models for ResNet and considering different pooling strategies
The Patel trials: further evidence of the need to reform the Griffith Codes
No full textThis article argues that the two trials of Dr Jayant Patel for criminal medical negligence under s 288 of the Criminal Code 1899 Act (Qld) highlight the inadequacies of the duty provisions in the Griffith Codes of Queensland and Western Australia. The difficulties with these duty provisions extend beyond causation and go to the heart of the construction of the Griffith Codes. The fundamental problem lies in the wording of s 23 of both the Queensland and
the Western Australia Codes, the principal section dealing with criminal responsibility, which allows a prosecution for criminal negligence under two alternative routes with different standards of proof, and the importation of
common law criminal negligence into the duty provisions in the absence of a specified fault element in the relevant Code sections. It is further contended that other criminal law jurisdictions in Australia, such as the Criminal Code
1995 (Cth), offer a better model for the prosecution of criminal negligence cases that flow from breach of a specified duty. The article has greatly benefited from comments provided to the author by Justice HG Fryberg, who
conducted the second Patel trial
Novel Motor-Shaped Rotational Inductor for Motor Drive Applications
No full textThis paper presents a validation of the novel motor-shaped rotational inductor. To validate the concept, 12 slots 2 poles rotational inductor is tested at different supply frequencies and rotor speeds. Experimental results have shown that the iron losses reduce as the rotor speed increases to the synchronous speed of the stator supply. The performance of the integrated rotational inductor is also compared with traditional EE core inductor in terms of total losses, synchronous inductance, copper resistance, and total harmonic distortion (THD). The total loss-to inductance ratio of the rotational inductor is reduced by 22.5% when rotor is rotating at 18 kRPM and supply frequency is held at 300 Hz. A significant reduction in copper resistance-to-inductance is also noticed when supply frequency is varied from 0 Hz to 20 kHz. Furthermore, the synchronous inductance and voltage and current's %THD of rotational inductor is found to be superior to EE core inductor
- …
