Earticle

Anomaly recognition in visual and audio data has gained increasing significance in computer vision, as it plays a crucial role in protecting human lives and property. In this work, we developed a semi-supervised multimodal framework for anomaly recognition that combines audio and visual data for better performance. The proposed framework employs a hybrid network consisting of a convolutional neural network, Bi-Directional Long Short-Term Memory, a multi-head attention module, and a fully connected layer for anomalous pattern recognition. We created a novel real-time visual-audio anomaly recognition dataset and evaluated our framework on it, achieving promising results.

Recently, a graph neural network has played a crucial role across various fields. In this paper, we designed a Graph Convolutional Network (GCN) to analyze population movement at the city level. It consists of four Graph Convolution (GC) layers, with each layer responsible for aggregating knowledge from its neighboring nodes and updating the feature representation for each city. We utilized population mobility data from China, which includes daily city-to-city movement data. GCN estimates the strength of relationships among all cities. Experimental results demonstrate that the proposed GCN achieves improved performance in estimating city-to-city migration flow relationships.

The identification of key feature selection plays a significant role in accurate population projection, which is an essential aspect of demographic statistics. The goal of this paper is to investigate the importance of the different features in population projection by using four advanced feature analysis techniques i.e. Canonical Correlation Analysis (CCA), Linear Discriminant Analysis (LDA) Fast Independent Component Analysis (FICA) and Principal Component Analysis (PCA). This analysis is important to determine the major factors that affect population change. The identification and ranking of these predictors can enhance demographic forecasting and policy planning. We utilized Koran population data from the UN Population Division dataset and evaluated the above four methods. The experimental results reveal that LDA achieved the lowest performance in selecting the most appropriate features, while PCA is the most efficient in selecting an effective feature with the highest variance. These insights build up the knowledge of population change and refine the projection models.

The battery market is experiencing rapid growth due to advancements in technology and increased recycling efforts. Verifying the suitability of developed batteries through rate capability experiments, which measure capacity based on charging and discharging speeds, is essential but resource-intensive and time-consuming. This research proposes a method to predict battery rate capability using a polynomial regression model based on similar data groups, aiming to shorten these experiments. The research was conducted in two main stages, namely the construction of the dataset and the development of the predictive model. Data was collected from experimental graphs in existing literature and new experiments on Coin Cell batteries. Through preprocessing steps including deduplication, interpolation, and extrapolation, a comprehensive dataset was created. A combined Quadratic and Linear Piecewise Interpolation method was developed to handle missing data efficiently. In the model development stage, polynomial regression models were created for groups of similar battery data, allowing accurate predictions for partial rate capability experiments. Experimental results demonstrated high accuracy, significantly reducing the need for extensive testing. The proposed method offers substantial time and resource savings, enhancing the efficiency of the battery development process.

In recent years, anomaly recognition using audio has attracted the attention of the research community, due to the increasing number of abnormal situations day by day. In the past, researchers have mainly focused on video-based anomaly recognition. However, occlusion is one of the most important factors due to which the anomalous object is unidentifiable. Therefore, in this paper, we proposed a modified vision transformer that utilized the Shifted Patch Tokenization (SPT), and Local Self-Attention (LSA) mechanism and reduced the number of multilayer perceptrons in the head, enabling the model to capture rich spatial information within the spectrogram of anomalous data. The proposed model is implemented using the Sound Events for Surveillance Applications (SESA) dataset and obtained 87% testing accuracy. Thus, the proposed model is an efficient and effective solution for audio-based anomaly recognition.

In recent years, Generative Adversarial Networks (GANs) appeared as a prevailing solution for combating data scarcity in various domains. This study delves into utilizing WaveGAN, a specialized GAN architecture, to address the inherent challenges stemming from the limited availability of audio datasets. Our primary objective is to tackle the issue of constrained audio data resources by utilizing the potential of WaveGAN. Our research is driven by the overarching goal of investigating the capacity of CNN to gather significant insights from an extensive corpus of human speech data. A key focus of our work is to demonstrate the effectiveness of WaveGAN in generating synthetic audio data, thereby expanding the breadth of our audio dataset and bolstering the resilience of our classification models. Our study aims to yield improved classification results, providing crucial insights into the viability of this approach in alleviating data scarcity challenges of audio analysis.

This paper introduces a comprehensive approach to dataset standardization aimed at enhancing the effectiveness and reliability of solar power forecasting models. Leveraging multiple datasets, this study incorporates additional attributes such as atmospheric pressure and sunshine duration. These enrichments bridge critical gaps in meteorological and environmental data, facilitating more robust and precise solar power forecasting. The paper underscores the significance of these attributes, furnishes detailed equations for their computation, and presents the outcomes of their integration. It underscores their pivotal role in enabling solar energy stakeholders to make informed decisions and optimize energy production effectively.

Fire detection is a significant attempt for preserving public safety in complex surveillance environments. Although advances in deep learning for fire detection, the task remains challenging due to the natural irregularity in fire images, including differences in lighting conditions, occlusions, and background complexity. To address these challenges, we present a novel framework for fire detection named fire channel attention network (FCAN), which is capable of differentiating challenging fire scenes. Our approach is motivated by the need to enhance the accuracy of fire detection by selectively emphasizing the most informative channels of the input image through a channel attention (CA). Furthermore, our model captures the salient features from the input image and suppresses the irrelevant ones, thereby overcoming the aforementioned challenges of fire detection. The FCAN is evaluated on two benchmark datasets and surpassed existing methods in terms of accuracy and F1 score. The proposed model demonstrates the effectiveness of fire detection, highlighting its potential for practical applications in fire safety and prevention.

The analysis, recognition and perception of behavior has usually been a crucial task for researchers. The goal of this paper is to address the problem to recognize animal species, which has numerous applications in zoology, ecology, biology, and entertainment. Researchers used different machine learning approach for animal species recognition, however the researchers mostly used image data for this purpose and ignore the importance of audio data. In this work, our focus is to process multi modality (image and voice) dataset for animal species recognition. We proposed two different networks for animals’ audio and visual representation to recongize animals’ species. First network for animals’ audios classification that extract MFCC features, and these features is passed from four VGG style blocks while the second network extract visual features from images to classify according to their species. The experimental results demonstrated the effectiveness of the proposed model of achieved better performance in terms of classification accuracies.

Deep learning models achieved a lot of success due to the availability of labeled training data. In contrast, labeling a huge amount of data by a human is a time-consuming and expensive solution. Active Learning (AL) efficiently addresses the issue of labeled data collection at a low cost by picking the most useful samples from a large number of unlabeled datasets However, current AL techniques largely depend on regular human involvement to annotate the most uncertain/informative samples in the collection. Therefore, a novel AL-based framework is proposed comprised of proxy and active models to reduce the manual labeling costs. In the proxy model, VGG-16 is trained on chunks of labeled data that later act as an annotator decision. On the other hand, in the active model, unlabeled is passed to Inception-V3 using the sampling strategy. The uncorrected predicted samples are then forwarded to the proxy model for annotation and considered those data have a high confidence score. The empirical results verify that our proposed model is the best in terms of annotation and accuracy.

Research on fire detection has grown steadily over the last few decades which is the key concern of the research community to prevent the lives of mankind and their property from damages. Several researchers developed video fire detection datasets and proposed different machine learning algorithms for its accurate detection. Therefore, it is very significant for the researchers to understand the relevant datasets in this field that can provide help in terms of results comparison and speed up the research based on the existing datasets instead of creating a new dataset. In this paper, we provide a comprehensive overview of existing fire detection datasets. Firstly, we reviewed seven different fire detection datasets in detail, which can provide helps to new researchers in this field. Secondly, we provided a detailed description of these datasets, and analyzed the shortcomings and suggestions for further fruitful research. This paper is helpful for new researchers to identify possible research gaps and limitations about fire detection datasets.

Fire is an extremely catastrophic disaster that leads to the destruction of forests, human assets, reduced soil fertility, land resources, and the cause of global warming. In the current decade, fire detection and its management are the major concern of several researchers to prevent social, ecological, and economic damages. To overcome such kind of losses, early fire detection, and the automatic response is very significant. Moreover, achieving high accuracy with reducing inference time and model size is also challenging for the Unmanned Aerial Vehicle (UAVs). Therefore, in this work, we enabled the VGG16 architecture for UAV in terms of reducing its learning parameters from 138 million to 11.4 million for early fire detection. The proposed system is inexpensive in terms of computation and size. The performance of our proposed work is evaluated over the custom dataset. We performed comprehensive experiments using various deep learning architectures such as VGG16, ResNet50, and the proposed CNN model. The experimental results based on the proposed model achieved an accuracy of 98% on 50 epochs.

오늘날 현대사회에서 스마트하고 안전한 도시는 연구 커뮤니티의 주요 관심사 중 하나이다. 도시들은 개방된 지역, 농경지, 숲으로 둘러싸여 있으며, 화재 발생은 인간의 삶을 위협하고 그들의 재산도 손상시킬 수 있다. 최근 비전 센서 기반 화재 감지 기술은 컴퓨터 비전 분야의 전문가들을 통해, 최신 문헌에서 다양한 컨볼루션 신경 네트워크 (CNN)을 대한 최고의 성능을 달성하고 있다. 그러나 이러한 기술은 변환 불변이고, 지역성에 민감하며, 이미지에 대한 전체적인 이해가 부족하다. 또한 CNN 기반 모델은 계산 비용을 줄이기 위해 차원 축소를 위한 풀링 레이어 전략을 사용했지만, 가장 활동적인 특징 검출기의 정확한 위치와 같은 많은 의미 있는 정보를 손실한다. 이러한 문 제를 극복하기 위해 본 연구에서는 비전 트랜스포머(ViT)기반 화재 감지 모델을 개발하였다. ViT는 입력 이미지를 이미지 패치로 분할한 다음 워드 임베딩과 유사한 시퀀스 구조로 트랜스포머에 제공한다. 우리는 벤치마크 화재 데 이터 세트에서 제안된 작업의 성능을 평가하고 최신(SOTA) CNN 방법과 비교할 때 좋은 결과를 달성한다.

In today's modern age, smart and safe cities are one of the major concerns of the research community. The cities are surrounded by open areas, agricultural land, and forests, where fire incidence can make human lives threatening, damaging their properties as well. Recently, vision sensors-based fire detection has attracted computer vision domain experts, where the leading performance is achieved by a variety of convolution neural networks(CNN) in the recent literature. However, these techniques are translation invariant, locality-sensitive, and lacking a global understanding of images. Furthermore, CNN-based models use the pooling layers strategy for dimensionality reduction to reduce the computational cost but it also loses a lot of meaningful information such as the precise location of the most active feature detector. To overcome these problems, in this work, we developed Vision Transformers(ViT) based model for fire detection. The ViT split the input image into image patches and then feed these patches to the transformer in a sequence structure similar to word embeddings. We evaluate the performance of the proposed work on the benchmark fire dataset and achieve good results when compared to state-of-the-art(SOTA) fire detection CNN models.

Forest fire is one of the most dangerous disasters worldwide, due to which its management is a key concern of the research community to prevent social, ecological, and economic damages. Wildfires are extremely catastrophic disasters that lead to the destruction of forests, human assets, reduction of soil fertility and cause global warming. To overcome such kind of losses early fire detection and quick response is the key concern of research community. Therefore, in this paper, we propose a lightweight convolution neural network (CNN) method to efficiently detect the forest fire for unmanned aerial vehicles (UAVs) or drones. For the experimental evaluations, we develop an aerial images dataset from YouTube, movies, and google images. The results of the proposed architecture reveal its good performance in terms of 96% accuracy.