Aniket Agarwal | V.I. Vernadsky Crimean Federal University (original) (raw)

Papers by Aniket Agarwal

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

generalizes well on many VRR benchmarks. Our model outperforms the best-performing models on two ... more generalizes well on many VRR benchmarks. Our model outperforms the best-performing models on two large-scale long-tail VRR benchmarks, VG8K-LT (+2.0% overall acc) and GQA-LT (+26.0% overall acc), both having a highly skewed distribution towards the tail. It also achieves strong results on the VG200 relation detection task. Our code is available at https://github.com/Vision-CAIR/ RelTransformer.

INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT

The COVID-19 is an uncommon emergency that has brought about an enormous number of setbacks and s... more The COVID-19 is an uncommon emergency that has brought about an enormous number of setbacks and security issues. Individuals habitually wear covers to safeguard themselves from the spread of Covid. Since specific pieces of the face are covered up, face acknowledgment turns out to be incredibly troublesome. During the progressing Covid pandemic, one of the specialists' essential objectives is to concocted answers for the issue that are both fast and effective. The reason for this paper is to give a survey of different strategies and calculations for human acknowledgment with a facial covering. This paper depicts different methodologies, for example, Haar course, Adaboost, VGG-16 CNN Model, etc. An examination of these techniques is acted to figure out which approach is achievable. With innovative progressions. Keywords — Viola Jones, Adaboost, Computer Vision, Convolutional Neural Network, MobileNetV2, VGG – 16 Model.

2021 IEEE/CVF International Conference on Computer Vision (ICCV), 2021

Several approaches have been proposed in recent literature to alleviate the long-tail problem, ma... more Several approaches have been proposed in recent literature to alleviate the long-tail problem, mainly in object classification tasks. In this paper, we make the first largescale study concerning the task of Long-Tail Visual Relationship Recognition (LTVRR). LTVRR aims at improving the learning of structured visual relationships that come from the long-tail (e.g., "rabbit grazing on grass"). In this setup, the subject, relation, and object classes each follow a long-tail distribution. To begin our study and make a future benchmark for the community, we introduce two LTVRRrelated benchmarks, dubbed VG8K-LT and GQA-LT, built upon the widely used Visual Genome and GQA datasets. We use these benchmarks to study the performance of several state-of-the-art long-tail models on the LTVRR setup. Lastly, we propose a visiolinguistic hubless (VilHub) loss and a Mixup augmentation technique adapted to LTVRR setup, dubbed as RelMix. Both VilHub and RelMix can be easily integrated on top of existing models and despite being simple, our results show that they can remarkably improve the performance, especially on tail classes. Benchmarks, code, and models have been made available at: https://github.com/Vision-CAIR/LTVRR.

Egyptian Journal of Radiology and Nuclear Medicine, 2022

Background Coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome corona... more Background Coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which was declared a pandemic by the World Health Organization on 11 March 2020 has been reported in most countries around the world since its origins in Wuhan, China. As of September 2021, there have been over 229 million cases of COVID-19 reported worldwide, with over 4.7 million COVID-19–associated deaths. Body The devastating second wave of the COVID-19 pandemic in India has seen a rise in various extrapulmonary manifestations. One of key components in the pathogenesis of COVID-19 is downregulation of ACE-2, which is expressed on many organs and counterbalances the pro-inflammatory effects of ACE/angiotensin-II axis. This leads to influx of inflammatory cells into alveoli, increased vascular permeability and activation of prothrombotic mediators. Imaging findings such as ground glass opacities, interlobular septal thickening, vascular dilatation and pulmonary th...

International Journal of System Dynamics Applications, 2021

The main objective of the paper is to minimize the use of conventional generators and optimize th... more The main objective of the paper is to minimize the use of conventional generators and optimize the fuel cost. To minimize the use of conventional generators, solar thermal power plant (STPP) is proposed in this paper. An approach for optimal location of STPP is also proposed in this paper. To minimize the fuel cost, firstly unit commitment (UC) is applied in conventional generators. Then genetic algorithm (GA) is used to optimize the fuel cost of committed generators. The suggested method is tested on an IEEE 14 bus test system for 24 hr. schedule with variable load. The effectiveness of the proposed methodology is illustrated in three cases. Case 1 is used to identify the STPP location to reduce the fuel cost of conventional generator. In Case 2, unit-commitment is applied to save considerable fuel input and cost. In order to optimize the committed fuel cost, a genetic algorithm is applied in Case 3.

Lecture Notes in Electrical Engineering, 2021

Currently, the majority of the world's electricity demand is met by thermal power generation stat... more Currently, the majority of the world's electricity demand is met by thermal power generation stations that run purely on traditional fossil fuels. Utilities rely mainly on these sources to spend huge revenue to meet the ever-increasing demand for electricity. This motivates utilities to manage their generation most cost-effectively according to the load demand. Thus, an optimum generation allocation among the various power generating units can save considerable fuel inputs and expenses. Extending this optimization technique to decide which of these units would participate in the optimum allocation could theoretically save a greater amount of fuel costs. In other words, the determination of whether the device has to be ON/OFF is important. This is termed as unit commitment (UC). As for deciding the optimal generation dispatch for the minimum cost, a complete optimal power flow (OPF) is run over the UC time horizon for each hour's commitment. Conventional OPF solves all constraints such as fixed bus voltage limits, line power flows, transformer tap positions, etc., for optimum dispatch adjustment, resulting in a secure solution. In this paper, with the integration of solar thermal power plant, the total generation cost is reduced. The suggested method is tested by applying it to the standard IEEE 14 bus test system. The findings of the UC, demonstrated in both the presence and absence of STPP, show the method's efficiency. We propose a mathematical programming-based approach with alternative current optimal power flow (ACOPF) network constraints, to optimize the unit commitment problem.

ArXiv, 2021

Visual relationship recognition (VRR) is a fundamental scene understanding task. The structure th... more Visual relationship recognition (VRR) is a fundamental scene understanding task. The structure that VRR provides is essential to improve the AI interpretability in downstream tasks such as image captioning and visual question answering. Several recent studies showed that the long-tail problem in VRR is even more critical than that in object recognition due to the compositional complexity and structure. To overcome this limitation, we propose a novel transformerbased framework, dubbed as RelTransformer, which performs relationship prediction using rich semantic features from multiple image levels. We assume that more abundant contextual features can generate more accurate and discriminative relationships, which can be useful when sufficient training data are lacking. The key feature of our model is its ability to aggregate three different-level features (local context, scene, and dataset-level) to compositionally predict the visual relationship. We evaluate our model on the visual ge...

ArXiv, 2020

Understanding a scene by decoding the visual relationships depicted in an image has been a long s... more Understanding a scene by decoding the visual relationships depicted in an image has been a long studied problem. While the recent advances in deep learning and the usage of deep neural networks have achieved near human accuracy on many tasks, there still exists a pretty big gap between human and machine level performance when it comes to various visual relationship detection tasks. Developing on earlier tasks like object recognition, segmentation and captioning which focused on a relatively coarser image understanding, newer tasks have been introduced recently to deal with a finer level of image understanding. A Scene Graph is one such technique to better represent a scene and the various relationships present in it. With its wide number of applications in various tasks like Visual Question Answering, Semantic Image Retrieval, Image Generation, among many others, it has proved to be a useful tool for deeper and better visual relationship understanding. In this paper, we present a de...

ArXiv, 2019

In this work, we study the problem of training deep networks for semantic image segmentation usin... more In this work, we study the problem of training deep networks for semantic image segmentation using only a fraction of annotated images, which may significantly reduce human annotation efforts. Particularly, we propose a strategy that exploits the unpaired image style transfer capabilities of CycleGAN in semi-supervised segmentation. Unlike recent works using adversarial learning for semi-supervised segmentation, we enforce cycle consistency to learn a bidirectional mapping between unpaired images and segmentation masks. This adds an unsupervised regularization effect that boosts the segmentation performance when annotated data is limited. Experiments on three different public segmentation benchmarks (PASCAL VOC 2012, Cityscapes and ACDC) demonstrate the effectiveness of the proposed method. The proposed model achieves 2-4% of improvement with respect to the baseline and outperforms recent approaches for this task, particularly in low labeled data regime.

arXiv: Computer Vision and Pattern Recognition, 2020

Several approaches have been proposed in recent literature to alleviate the long-tail problem, mo... more Several approaches have been proposed in recent literature to alleviate the long-tail problem, mostly in the object classification task. We propose to study the task of Long-Tail Visual Relationship Recognition (LTVRR), which aims at generalizing on the structured long-tail distribution of visual relationships (e.g., "rabbit grazing on grass"). In this setup, subject, relation, and object classes individually follow a long-tail distribution. We first introduce two large-scale long-tail visual relationship recognition benchmarks to study this task, dubbed as VG8K-LT (5330 objects, 2000 relationships) and GQA-LT (1703 objects, 310 relations). VG8K-LT and GQA-LT are built upon the widely used Visual Genome and GQA datasets. In contrast to existing benchmarks, some classes appear at a very low frequency ($1-14$ examples). We use these benchmarks to study the performance of several state-of-the-art long-tail models on LTVRR setup. We developed a visiolinguistic hubless (ViLHub)...

2020 International Conference on Electrical and Electronics Engineering (ICE3), 2020

Smart cities have increased values as they are solution to the various challenges that modern cit... more Smart cities have increased values as they are solution to the various challenges that modern cities faces from the last decades. The concept of smart city is an intellectual approach towards sustainable development and also provides solutions to the concerns like environment, quality of life and others. The major problems in the Indian cities are that of water and electricity, hence there is need to conserve water. Electric energy sources are of such a nature that they do not affect the environment. This paper proposes basic components of smart city like the smart infrastructure, smart water management, smart electric energy management, smart waste management and smart industrial environment and their hardware implementation is shown to achieve sustainable development.

Egyptian Journal of Radiology and Nuclear Medicine, 2021

Background Objectives: To comparatively evaluate the role of ultrasound and MRI in rotator cuff a... more Background Objectives: To comparatively evaluate the role of ultrasound and MRI in rotator cuff and biceps tendon pathologies and to establish ultrasound as a consistently reproducible, quick and accurate primary investigation modality sufficient to triage patients requiring surgical correction of full thickness rotator cuff tears. Methods: Fifty patients, clinically suspected to have rotator cuff and/or biceps tendon pathologies, with no contraindications to MRI, were evaluated by US and MRI, in a prospective cross-sectional observational study. US was done with high-frequency linear probe, and MRI was done on a 1.5-T scanner using T1 oblique sagittal, proton density (PD)/T2 fat-suppressed (FS) oblique sagittal, T1 axial, PD/T2 FS axial, T1 oblique coronal, T2 oblique coronal and PD FS oblique coronal sequences. Statistical testing was conducted with the statistical package for the social science system version SPSS 17.0. The sensitivity, specificity, PPV, NPV and accuracy were als...

Indian Journal of Case Reports, 2021

C oronavirus disease 2019 (COVID-19) is an infectious disease caused by severe acute respiratory ... more C oronavirus disease 2019 (COVID-19) is an infectious disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), a strain of coronavirus. The outbreak was officially recognized as a pandemic by the World Health Organization (WHO) on March 11, 2020. The patients with COVID-19 present primarily with fever, myalgia or fatigue, and dry cough. Although most patients have a favorable prognosis, older patients and those with chronic underlying conditions such as diabetes and hypertension may have poor outcomes [1]. The patients with severe acute respiratory illness (SARI) may develop dyspnea and hypoxemia within 1 week after the onset of the disease, which may quickly progress to acute respiratory distress syndrome (ARDS) with or without other end-organ failures [2]. A definitive diagnosis of COVID-19 is by reverse transcriptase polymerase chain reaction (RT-PCR) test and computed tomography (CT) scan is an adjunct modality helpful in assessing complications. We report a case of clinically suspected patient with COVID-19 who developed ARDS and pulmonary thrombosis after hospital admission and was lost on the 14th day of hospital admission. CASE REPORT A 53-year-old man, diagnosed case of hypertension and diabetes mellitus for 14 years, presented with history of undocumented fever, malaise, occasional dry cough, and breathlessness for 5 days in the month of May 2020 at the peak of the pandemic in our city. The general condition of the patient was poor. At the time of admission, the blood pressure was 152/96, respiratory rate was 20/min, and the temperature was 102.4°C. The patient was hypoxic at admission (PaO 2 of 60.6 and PaCO 2 of 52.2 with a pH of 7.145, and serum HCO 3 of 18.2). Blood sugars were uncontrolled at admission (HbA1c of 18). He also had a derangement of renal functions (creatinine of 2.9 with an estimated glomerular filtration rate of 19.6). Chest radiograph done on the day of admission revealed a right parahilar consolidation and a few patchy opacities in both lung fields (Fig. 1). Since the patient developed hemoptysis and continued to be hypoxic, a clinical diagnosis of pulmonary thromboembolism was considered. D-dimer was raised four times. Doppler of bilateral lower limbs revealed no evidence of deep venous thrombosis (DVT). CT pulmonary angiography (CTPA) was done on day 5 for confirmation. CTPA revealed thrombosis in multiple segmental branches in both lung fields (right > left) (Fig. 2a-h). A dumbbellshaped aneurysm measuring 2.8 × 2 cm was also noted arising ABSTRACT COVID-19 caused by the SARS-CoV-2 strain of coronavirus was officially recognized as a pandemic by the World Health Organization on March 11, 2020. Most patients present with mild disease. However, elderly patients and those with coexisting comorbid conditions such as diabetes and hypertension are more likely to develop severe disease. Definitive diagnosis is by RT-PCR test and CT scan is an adjunct modality. By virtue of being a hypercoagulable state with cytokine storm and microthrombosis as key components in pathogenesis, additional finding of pulmonary thrombosis in such patients should increase diagnostic accuracy. We report an interesting case with clinical and radiological features supporting COVID-19 pneumonia such as patchy ground-glass opacities with consolidation in bilateral peripheral lung fields along with segmental pulmonary thrombosis. The conundrum in the case arises from the negative RT-PCR test and presence of pulmonary artery aneurysm which could be an incidental finding or sequelae of COVID-19, which remains to be studied.

Artificial Intelligence Review, 2019

The combined impact of new computing resources and techniques with an increasing avalanche of lar... more The combined impact of new computing resources and techniques with an increasing avalanche of large datasets, is transforming many research areas and may lead to technological breakthroughs that can be used by billions of people. In the recent years, Machine Learning and especially its subfield Deep Learning have seen impressive advances. Techniques developed within these two fields are now able to analyze and learn from huge amounts of real world examples in a disparate formats. While the number of Machine Learning algorithms is extensive and growing, their implementations through frameworks and libraries is also extensive and growing too. The software development in this field is fast paced with a large number of open-source software coming from the academy, industry, start-ups or wider open-source communities. This survey presents a recent time-slide comprehensive overview with comparisons as well as trends in development and usage of cutting-edge Artificial Intelligence software. It also provides an overview of massive parallelism support that is capable of scaling computation effectively and efficiently in the era of Big Data.

African Journal of Urology

Background Myelolipomas are mesenchymal tumors usually involving the adrenal gland. They are a ra... more Background Myelolipomas are mesenchymal tumors usually involving the adrenal gland. They are a rare entity with incidence ranging from 0.08 to 0.4% as per autopsy reports. Only 15% of these are in extra-adrenal locations such as pelvis, retroperitoneum, mediastinum, lungs, stomach, liver, spleen and kidneys. Very few (only 3) cases of bilateral extra-adrenal perirenal myelolipomas have been reported until now; and one such case has been presented in our case report. Case presentation A 45-year-old male presented with abdominal distension and bilateral lower limb edema for 2 months. Renal functions were mildly deranged. Ultrasound was suggestive of perirenal masses. CT scan of the abdomen confirmed the diagnosis of bilateral extra-adrenal, perirenal myelolipoma. The differentials of lipomatous retroperitoneal tumors were considered. Core biopsy from perirenal masses revealed adipose tissue with interspersed hematopoietic precursors. Conclusion Extra-adrenal myelolipoma are rare tumor...