Posting Techniques in Indoor Environments Based on Deep Learning for Intelligent Building Lighting System

Abstract

Recently, with the rapid development of society, solutions to reduce energy consumption in the world have attracted a lot of attention, especial electric energy. In this regard, a system that can control light on and off by determining the location of the person to reduce the waste of electricity used in public buildings, called intelligent building lighting system. Following the practical requirements of the intelligent building lighting system, a technique for positioning in indoor environments is proposed, supporting the design of a positioning system based on deep learning and the Cerebellar Model Articulation Controller (CMAC), called Y-CMAC.This technique adopts YOLOv3 (the method in the paper of YOLOv3: An Incremental Improvement) for object detections and makes the coordinate of a person in the image. On the other hand, using CMAC to calculate the actual position of the person in the indoor environment. Moreover, massive surveillance video is used to reduce the cost of equipment and facilitate the promotion of applications. The average positioning error is controlled at around 1m in this paper. INDEX TERMS Intelligent building lighting system, indoor positioning, monitor identification, CMAC, deep learning.

Automatic body region localization in 3D-CT images based on the improved YOLO model

“…More recently, there had been different versions of YOLO-model and their applications, e.g. YOLO-v1 model [12,13,14] , YOLO-v2 model [15,16,17] , and YOLO-v3 model [18,19] et al From the point of view of network architecture, the YOLO-v1 model contained 24 convolutional layers and two fully connected layers. The YOLO-v2 model removed the fully connected layers, but added a batch normalization behind each convolutional layer and performs normalization preprocessing for each batch of data.…”