.png){kind=icon}
.png){kind=icon}
Return
Indoor objects detection system implementation using multi-graphic processing units
.png){kind=icon}
PRE
AI
DOI：10.1007/s10586-021-03419-9.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Indoor objects detection and recognition plays an important role in computer science and artificial intelligence fields. This task plays also a crucial role especially for blind and visually impaired persons (VIP) assistance navigation. Aiming to address this problem, we propose in this paper to develop a new indoor object detection system based on deep learning algorithms. Unfortunately, this type of algorithms requires heavy computational resources, and energy consumption. To address this problem, we propose a CUDA multi-GPU framework implementation of the proposed system. Generally deep learning based algorithms require huge amount of data to train and test networks. We propose to develop a new indoor dataset which consists of 11,000 indoor images containing 25 indoor landmark objects highly recommended for blind and VIP navigation. Based on the obtained results, the developed system shows big efficiency in terms of detection accuracy as well as processing time.
Keywords:
Parallel computing
Deep learning
GPU implementation
Assistance navigation
Deep convolutional neural network (DCNNs)
Indoor object detection
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
C
IF:
4.1
Papers:
5.0K
Citations:
7.5K