In the field of object detection, recently, tremendous success is achieved, but still it is a very challenging task to detect and identify objects accurately with fast speed. Human beings can detect and recognize multiple objects in images or videos with ease regardless of the object’s appearance, but for computers it is challenging to identify and distinguish between things. In this paper, a modified YOLOv1 based neural network is proposed for object detection. The new neural network model has been improved in the following ways. Firstly, modification is made to the loss function of the YOLOv1 network. The improved model replaces the margin style with proportion style. Compared to the old loss function, the new is more flexible and more reasonable in optimizing the network error. Secondly, a spatial pyramid pooling layer is added; thirdly, an inception model with a convolution kernel of 1 ∗ 1 is added, which reduced the number of weight parameters of the layers. Extensive experiments on Pascal VOC datasets 2007/2012 showed that the proposed method achieved better performance.
Purpose Motion capture system (MoCap) has been used in measuring the human body segments in several applications including film special effects, health care, outer-space and under-water navigation systems, sea-water exploration pursuits, human machine interaction and learning software to help teachers of sign language. The purpose of this paper is to help the researchers to select specific MoCap system for various applications and the development of new algorithms related to upper limb motion. Design/methodology/approach This paper provides an overview of different sensors used in MoCap and techniques used for estimating human upper limb motion. Findings The existing MoCaps suffer from several issues depending on the type of MoCap used. These issues include drifting and placement of Inertial sensors, occlusion and jitters in Kinect, noise in electromyography signals and the requirement of a well-structured, calibrated environment and time-consuming task of placing markers in multiple camera systems. Originality/value This paper outlines the issues and challenges in MoCaps for measuring human upper limb motion and provides an overview on the techniques to overcome these issues and challenges.
In recent years, due to technological advancements, the concept of Industry 4.0 (I4.0) is gaining popularity, while presenting several technical challenges being tackled by both the industrial and academic research communities. Semantic Web including Knowledge Graphs is a promising technology that can play a significant role in realizing I4.0 implementations. This paper surveys the use of the Semantic Web and Knowledge Graphs for I4.0 from different perspectives such as managing information related to equipment maintenance, resource optimization, and the provision of on-time and on-demand production and services. Moreover, to solve the challenges of limited depth and expressiveness in the current ontologies, we have proposed an enhanced reference generalized ontological model (RGOM) based on Reference Architecture Model for I4.0 (RAMI 4.0). RGOM can facilitate a range of I4.0 concepts including improved asset monitoring, production enhancement, reconfiguration of resources, process optimizations, product orders and deliveries, and the life cycle of products. Our proposed RGOM can be used to generate a knowledge graph capable of providing answers in response to any real-time query.
No abstract
Flash floods and hurricanes are caused by the release of energy inside the oceans. Hurricanes are very sudden and may lead to heavy infrastructural damage with loss revenues associated human and animal's fatalities. Diversified techniques have been utilized to properly investigate the flash floods and hurricanes before the event. A hydro atmospheric and climatic change due to the hurricanes leads towards the high death toll. Approaches for the early prediction of flash floods and hurricanes may be categorized as (a) Modeling of the system (bathymetry), (b) Sensors and gauges-based measurement, (c) Radar-based images, (d) Satellite images and data, and (e) AI-based prediction. Comparative analysis of direct real-time data from the sensors and gauges, is more reliable compared to other techniques but it may contain some errors and missing information which leads towards the false alarms. Therefore, in this paper, a novel predictive hybrid algorithm (ANN PSO) has been applied to estimate the flash floods and hurricanes more precisely. A suitable combination of the sensors will give the benefit of better precision and improved accuracy when compare to the use of a single sensor. The combination of six process variables utilized in this paper for the measurement and investigation of the flash flood has been discussed. Real-time data of over forty eight (48) hours has been collected from PIR, Ultrasonic sensor, Temperature sensor, CO2 sensor, Rainfall module, Pressure, and temperature sensor. ANN feed-forward propagation is trained by using sample collected data from the multi-modal sensing device and applied for the classification of events while neurons are optimized by the particle swarm optimization (PSO), taking less processing time without requiring advanced complex computational resources. Results have proved that proposed AI based technique for the early identification of flash floods and hurricanes have worked more accurate and performance-wise better than the ongoing techniques. The results include flood probabilities and prediction analysis using proposed algorithm.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.