Redirigiendo al acceso original de articulo en 19 segundos...
Inicio  /  Information  /  Vol: 12 Par: 10 (2021)  /  Artículo
ARTÍCULO
TITULO

Towards Edge Computing Using Early-Exit Convolutional Neural Networks

Roberto G. Pacheco    
Kaylani Bochie    
Mateus S. Gilbert    
Rodrigo S. Couto and Miguel Elias M. Campista    

Resumen

In computer vision applications, mobile devices can transfer the inference of Convolutional Neural Networks (CNNs) to the cloud due to their computational restrictions. Nevertheless, besides introducing more network load concerning the cloud, this approach can make unfeasible applications that require low latency. A possible solution is to use CNNs with early exits at the network edge. These CNNs can pre-classify part of the samples in the intermediate layers based on a confidence criterion. Hence, the device sends to the cloud only samples that have not been satisfactorily classified. This work evaluates the performance of these CNNs at the computational edge, considering an object detection application. For this, we employ a MobiletNetV2 with early exits. The experiments show that the early classification can reduce the data load and the inference time without imposing losses to the application performance.

 Artículos similares

       
 
Chen Li, Hongming Wang and Peiting Sun    
It is well known that sail-assisted propulsion works under gradient wind conditions in the atmospheric boundary layer, and is an energy-saving device for fuel consumption. In order to study the aerodynamic characteristics of a wingsail in the atmospheric... ver más

 
Yusuf O. Busari, Yupiter H. P. Manurung, Martin Leitner, Yusuf L. Shuaib-Babata, Muhd F. Mat, Hassan K. Ibrahim, David Simunek and Mohd Shahar Sulaiman    
This research presents the numerical evaluation of fatigue crack growth of structural steels S355 and S960 based on Paris? law parameters (C and m) that are experimentally determined with a single edge notched tension (SENT) specimen using optical and cr... ver más
Revista: Applied Sciences

 
Bowen Li, Qiangqiang Sun, Dandan Xiao and Wenqiang Zhang    
The aerodynamics of an aerofoil with surface heating was numerically studied with the objective to build an effective anti-icing strategy and balance the aerodynamics performance and energy consumption. NACA0012, RAE2822 and ONERA M6 aerofoils were adopt... ver más
Revista: Aerospace

 
Maitrik Shah, Beena Baloni, Salim Channiwala     Pág. 216 - 227
The design and off-design performance of a centrifugal pump largely depends on geomechanical parameters. This study aims at enhancing the performance by optimizing three geomechanical parameters of impeller-volute interactions. The present optimization i... ver más

 
Maximilian Hoffmann and Ralph Bergmann    
Similarity-based retrieval of semantic graphs is a core task of Process-Oriented Case-Based Reasoning (POCBR) with applications in real-world scenarios, e.g., in smart manufacturing. The involved similarity computation is usually complex and time-consumi... ver más
Revista: Algorithms