SlideShare a Scribd company logo
1 of 27
Download to read offline
DeepRemote: A Smart Remote Controller for
Intuitive Control through Home Appliances
Recognition by Deep Learning
〇Yuta Takahashi1, Naoki Shirakura1, Kenta Toyoshima1, Takuro
Amako1, Ryota Isobe1, Jun Takamatsu1 and Keiichi Yasumoto1
1. Nara Institute of Science and Technology
The Tenth International Conference on Mobile
Computing and Ubiquitous Networking (ICMU2017)
S7: Application, October 5, 2017
Background
❖Increasing home appliances
2
❖Many IR remote controllers
▪ Need to learn how to use
▪ High management costs
Increase user’s burden
Unification of remote controllers
3
Network Multiple IR controller
Increase the number of
appliances
The interface becomes
complicated
Method for selecting home appliance is important!
✔ Management cost
Home appliance selection
4
❖With special attachments
❖Without special attachments
✔ Accurate selection
Cost of devices/markers
Selectable distance
Voice [Pan 2010] Vision [Kong 2016]
Intuitiveness ✔ Intuitiveness
IR control
Burden for wearing
• IR LED [Neßelrath 2011]
• IR transmitter [Tsukada 2004]
• QR [Ullah 2012]
Dedicated device is needed
DeepRemote (Proposed)
5
❖Home appliance selection
Object recognition by deep learning
→ Intuitive and robust
❖Two units
▪ Control unit
▪ Deep learning unit
❖Network
Home network (ROS)
Control unit
6
✓ Hand-held-type remote controller
❖Main processor
▪ Raspberry Pi3 (with Wi-Fi module)
❖Capturing a home appliance
▪ Front camera
❖Control interfaces
▪ Four buttons
▪ Gestures (right & left rotation)
Deep learning unit
7
✓ Image recognition
❖Main processor
▪ Laptop PC with Core i5 (Ubuntu)
▪ Distribute the calculation load of control unit
❖Recognition Model
▪ VGG16 [Simonyan 2014]
TV
Recognition
result
Deep
learning unit
Control unit
Technique of training
❖Requirements of deep learning with zero-base
8
Huge dataset A lot of time
❖Fine tuning
Trained model Arranging model
Re-training
Experiment
9
Experimental environment
10
✓Living room in smart-home facility
❖Five home appliances
▪ Fan (IR)
▪ Air conditioner (IR)
▪ TV (Network)
▪ Audio player (IR)
▪ Air purifier (Network)
❖Training data
▪ Took 20 images each appliances at P1, P2 and P3
▪ 20 × 5 (appliances) × 3 (positions) = 300 images
Example of training data
11
P1
P2
P3
Air
purifier
Air
conditioner
Audio
player Fan TV
Training
❖Model
▪ VGG16 trained ImageNet
❖Dataset
▪ Three hundred images
▪ Five classes
12
❖Optimizer
▪ SDG
❖Loss function
▪ Categorical cross entropy
Evaluations
1. Classification accuracy
▪ Captured 50 images of each appliance in each position
▪ 50 (images) × 5 (appliances) × 3 (positions) = 750 images
13
2. Response time
▪ Measured time since pushing button until return the result
▪ Evaluated at the same time as 1.
3. User test
▪ Verified control time of home appliances
Appliance Precision [%] Recall [%] F-measure [%]
P1
Air purifier 100.00 68.00 80.95
Audio player 83.33 90.00 86.54
TV 96.15 100.00 98.04
Air conditioner 57.47 100.00 72.99
Fan 86.96 40.00 54.80
Average 84.78 79.60 78.66
P2
Air purifier 81.63 80.00 80.81
Audio player 100.00 98.00 98.99
TV 92.59 100.00 96.15
Air conditioner 79.25 84.00 81.55
Fan 100.00 90.00 94.74
Average 90.69 90.40 90.45
P3
Air purifier 76.36 84.00 80.00
Audio player 100.00 58.00 73.42
TV 71.43 80.00 75.47
Air conditioner 53.17 84.00 65.12
Fan 100.00 62.00 76.54
Average 80.19 73.60 74.11
1. Classification accuracy of home appliances
14
P1:Effect of
black door?
P3: Too near?
Evaluations
1. Classification accuracy
▪ Captured 50 images of each appliance in each position
▪ 50 (images) × 5 (appliances) × 3 (positions) = 750 images
15
2. Response time
▪ Measured time since pushing button until return the result
▪ Evaluated at the same time as 1.
3. User test
▪ Verified control time of home appliances
2. Response time
16
(n=750)
Maximum time: 3.07 [sec]
Minimum time: 1.72 [sec]
Stable recognition about
two seconds
Evaluations
1. Classification accuracy
▪ Captured 50 images of each appliance in each position
▪ 50 (images) × 5 (appliances) × 3 (positions) = 750 images
17
2. Response time
▪ Measured time since pushing button until return the result
▪ Evaluated at the same time as 1.
3. User test
▪ Verified control time of home appliances
3. User test to assess control time (1/3)
Experimental conditions
Five participants
Position: P1
Targets: Fan, Air conditioner, TV and Audio player
18
❖Control time ❖Comparison
Holding Power on DeepRemote Original
vs
3. User test to assess control time (2/3)
19
F:86% F:98% F:77% F:55%
High → ✔ about 5 seconds
Lower than the 85% → over 10 seconds
Accuracy
Accuracy requires
85% over
3. User test to assess control time (3/3)
20
DeepRemote: Sum of each test > All
Original: Sum of each test < All
Changing time
DeepRemote < Original
Sum Sum
31.533.7
11.7
7.7
(- 2.1)
(+ 3.0)
Conclusions
❖DeepRemote
▪ Smart device for intuitively control the home appliances
▪ Deep learning method for home appliance selection
❖Results of evaluation
▪ 81.07% classification accuracy on average
▪ Average of response time is 1.97 seconds
▪ Time of power on an appliance takes 5 seconds
▪ 85% accuracy is required
▪ Time of changing target is lower than original remote
controller
21
Appendix
22
Future works
❖Improving recognition accuracy
▪ Cropping object
❖Investigating energy consumption
❖Accuracy of distinguishing similar appliances
❖Experiment in long-term usage
▪ User teaches correct label when the system recognizes
wrong
23
Detail conditions of user test
❖Compare DeepRemote and original remote controller
▪ measure control time
▪ Remove “air conditioner” (no original controller)
▪ Five participants (males in the 20s)
▪ Perform three times in each measurement (one participant performs
30 times)
▪ Participant’s position: P1
24
1) Measure one appliance control time
(start) → (power on an appliance)
2) Measure all appliances control time
(start) → (power on the fan)
→・ ・ ・ → (power on the audio player)
Confusion matrix of P1
25
Classified
Air purifier
Air
conditioner
Audio
player
Fan TV
True
Air purifier 34 6 9 0 1
Air
conditioner 0 50 0 0 0
Audio player 0 1 45 3 1
Fan 0 30 0 20 0
TV 0 0 0 0 50
Confusion matrix of P2
26
Classified
Air purifier
Air
conditioner
Audio
player
Fan TV
True
Air purifier 40 6 0 0 4
Air
conditioner 8 42 0 0 0
Audio player 1 0 49 0 0
Fan 0 5 0 45 0
TV 0 0 0 0 50
Confusion matrix of P3
27
Classified
Air purifier
Air
conditioner
Audio
player
Fan TV
True
Air purifier 42 5 0 0 3
Air
conditioner 7 42 0 0 1
Audio player 3 6 29 0 12
Fan 0 19 0 31 0
TV 3 7 0 0 40

More Related Content

What's hot

Norsonic nor131- 132-brochure
Norsonic nor131- 132-brochureNorsonic nor131- 132-brochure
Norsonic nor131- 132-brochurePhongNguyen515
 
Vibration Measurements In Wind Power Turbines
Vibration Measurements In Wind Power TurbinesVibration Measurements In Wind Power Turbines
Vibration Measurements In Wind Power Turbinesalexzio
 
The Minnich Auto Vibe System
The Minnich Auto Vibe SystemThe Minnich Auto Vibe System
The Minnich Auto Vibe SystemJill Reeves
 
Electronic units and systems for armored vehicles
Electronic units and systems for armored vehiclesElectronic units and systems for armored vehicles
Electronic units and systems for armored vehiclesAndrew Diakoniuk
 
Firelaser-datasheet-SenseTek Glasvezel detectie
Firelaser-datasheet-SenseTek  Glasvezel detectieFirelaser-datasheet-SenseTek  Glasvezel detectie
Firelaser-datasheet-SenseTek Glasvezel detectieHans Bronkhorst
 
Wind turbine vibration analysis
Wind turbine vibration analysisWind turbine vibration analysis
Wind turbine vibration analysisMartin Gascon
 
monitoring system Up guard 2001-en
monitoring system Up guard 2001-enmonitoring system Up guard 2001-en
monitoring system Up guard 2001-enFatih BATAL
 
Measuring the cutoff frequency of a low pass filter
Measuring the cutoff frequency of a low pass filterMeasuring the cutoff frequency of a low pass filter
Measuring the cutoff frequency of a low pass filterHasnain Ali
 
Fire alarm 3-e-public_emergency_alarm_reporting_systems
Fire alarm 3-e-public_emergency_alarm_reporting_systemsFire alarm 3-e-public_emergency_alarm_reporting_systems
Fire alarm 3-e-public_emergency_alarm_reporting_systemsgrantlerc
 
IR Controller For Air Conditioning Unit
IR Controller For Air Conditioning UnitIR Controller For Air Conditioning Unit
IR Controller For Air Conditioning Unittanvir26
 
Digital Recorder Presentation- Draft 1
Digital Recorder Presentation- Draft 1Digital Recorder Presentation- Draft 1
Digital Recorder Presentation- Draft 1Meshal Alawwad
 
Natural ventilation controls brochure
Natural ventilation controls brochureNatural ventilation controls brochure
Natural ventilation controls brochureAndy Brown
 
Fire alarm 3-d-supervising_station_alarm_system
Fire alarm 3-d-supervising_station_alarm_systemFire alarm 3-d-supervising_station_alarm_system
Fire alarm 3-d-supervising_station_alarm_systemgrantlerc
 
9041 Ultra II C trend 0714
9041 Ultra II C trend 07149041 Ultra II C trend 0714
9041 Ultra II C trend 0714Stephen Hawes
 
Up guard 2001-en
Up guard 2001-enUp guard 2001-en
Up guard 2001-enFatih BATAL
 
Harsha s ipmi_tool_osi
Harsha s ipmi_tool_osiHarsha s ipmi_tool_osi
Harsha s ipmi_tool_osisuniltomar04
 
1362576033 hcp measurement in neuropathy
1362576033 hcp measurement in  neuropathy 1362576033 hcp measurement in  neuropathy
1362576033 hcp measurement in neuropathy dfsimedia
 

What's hot (19)

Norsonic nor131- 132-brochure
Norsonic nor131- 132-brochureNorsonic nor131- 132-brochure
Norsonic nor131- 132-brochure
 
Vibration Measurements In Wind Power Turbines
Vibration Measurements In Wind Power TurbinesVibration Measurements In Wind Power Turbines
Vibration Measurements In Wind Power Turbines
 
The Minnich Auto Vibe System
The Minnich Auto Vibe SystemThe Minnich Auto Vibe System
The Minnich Auto Vibe System
 
Electronic units and systems for armored vehicles
Electronic units and systems for armored vehiclesElectronic units and systems for armored vehicles
Electronic units and systems for armored vehicles
 
Firelaser-datasheet-SenseTek Glasvezel detectie
Firelaser-datasheet-SenseTek  Glasvezel detectieFirelaser-datasheet-SenseTek  Glasvezel detectie
Firelaser-datasheet-SenseTek Glasvezel detectie
 
datalogger
dataloggerdatalogger
datalogger
 
Wind turbine vibration analysis
Wind turbine vibration analysisWind turbine vibration analysis
Wind turbine vibration analysis
 
monitoring system Up guard 2001-en
monitoring system Up guard 2001-enmonitoring system Up guard 2001-en
monitoring system Up guard 2001-en
 
Measuring the cutoff frequency of a low pass filter
Measuring the cutoff frequency of a low pass filterMeasuring the cutoff frequency of a low pass filter
Measuring the cutoff frequency of a low pass filter
 
Dgs co 968-034 digital co sensor module
Dgs co 968-034 digital co sensor moduleDgs co 968-034 digital co sensor module
Dgs co 968-034 digital co sensor module
 
Fire alarm 3-e-public_emergency_alarm_reporting_systems
Fire alarm 3-e-public_emergency_alarm_reporting_systemsFire alarm 3-e-public_emergency_alarm_reporting_systems
Fire alarm 3-e-public_emergency_alarm_reporting_systems
 
IR Controller For Air Conditioning Unit
IR Controller For Air Conditioning UnitIR Controller For Air Conditioning Unit
IR Controller For Air Conditioning Unit
 
Digital Recorder Presentation- Draft 1
Digital Recorder Presentation- Draft 1Digital Recorder Presentation- Draft 1
Digital Recorder Presentation- Draft 1
 
Natural ventilation controls brochure
Natural ventilation controls brochureNatural ventilation controls brochure
Natural ventilation controls brochure
 
Fire alarm 3-d-supervising_station_alarm_system
Fire alarm 3-d-supervising_station_alarm_systemFire alarm 3-d-supervising_station_alarm_system
Fire alarm 3-d-supervising_station_alarm_system
 
9041 Ultra II C trend 0714
9041 Ultra II C trend 07149041 Ultra II C trend 0714
9041 Ultra II C trend 0714
 
Up guard 2001-en
Up guard 2001-enUp guard 2001-en
Up guard 2001-en
 
Harsha s ipmi_tool_osi
Harsha s ipmi_tool_osiHarsha s ipmi_tool_osi
Harsha s ipmi_tool_osi
 
1362576033 hcp measurement in neuropathy
1362576033 hcp measurement in  neuropathy 1362576033 hcp measurement in  neuropathy
1362576033 hcp measurement in neuropathy
 

Viewers also liked

( (0) / (0)) ☆祝☆ (nari_ex) LT at Cloud Server Festa 2013 Autumn「サバフェス!」
( (0) / (0)) ☆祝☆ (nari_ex) LT at Cloud Server Festa 2013 Autumn「サバフェス!」( (0) / (0)) ☆祝☆ (nari_ex) LT at Cloud Server Festa 2013 Autumn「サバフェス!」
( (0) / (0)) ☆祝☆ (nari_ex) LT at Cloud Server Festa 2013 Autumn「サバフェス!」Narimichi Takamura
 
Yahoo! JAPANのサービス開発を10倍早くした社内PaaS構築の今とこれから
Yahoo! JAPANのサービス開発を10倍早くした社内PaaS構築の今とこれからYahoo! JAPANのサービス開発を10倍早くした社内PaaS構築の今とこれから
Yahoo! JAPANのサービス開発を10倍早くした社内PaaS構築の今とこれからYahoo!デベロッパーネットワーク
 
データテクノロジースペシャル:Yahoo! JAPANにおけるメタデータ管理の試み
データテクノロジースペシャル:Yahoo! JAPANにおけるメタデータ管理の試みデータテクノロジースペシャル:Yahoo! JAPANにおけるメタデータ管理の試み
データテクノロジースペシャル:Yahoo! JAPANにおけるメタデータ管理の試みYahoo!デベロッパーネットワーク
 
20171104 jaws festa 基調講演 配布用
20171104 jaws festa 基調講演 配布用20171104 jaws festa 基調講演 配布用
20171104 jaws festa 基調講演 配布用Kenji Tomooka
 
SVMについて
SVMについてSVMについて
SVMについてmknh1122
 
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017Carol Smith
 

Viewers also liked (6)

( (0) / (0)) ☆祝☆ (nari_ex) LT at Cloud Server Festa 2013 Autumn「サバフェス!」
( (0) / (0)) ☆祝☆ (nari_ex) LT at Cloud Server Festa 2013 Autumn「サバフェス!」( (0) / (0)) ☆祝☆ (nari_ex) LT at Cloud Server Festa 2013 Autumn「サバフェス!」
( (0) / (0)) ☆祝☆ (nari_ex) LT at Cloud Server Festa 2013 Autumn「サバフェス!」
 
Yahoo! JAPANのサービス開発を10倍早くした社内PaaS構築の今とこれから
Yahoo! JAPANのサービス開発を10倍早くした社内PaaS構築の今とこれからYahoo! JAPANのサービス開発を10倍早くした社内PaaS構築の今とこれから
Yahoo! JAPANのサービス開発を10倍早くした社内PaaS構築の今とこれから
 
データテクノロジースペシャル:Yahoo! JAPANにおけるメタデータ管理の試み
データテクノロジースペシャル:Yahoo! JAPANにおけるメタデータ管理の試みデータテクノロジースペシャル:Yahoo! JAPANにおけるメタデータ管理の試み
データテクノロジースペシャル:Yahoo! JAPANにおけるメタデータ管理の試み
 
20171104 jaws festa 基調講演 配布用
20171104 jaws festa 基調講演 配布用20171104 jaws festa 基調講演 配布用
20171104 jaws festa 基調講演 配布用
 
SVMについて
SVMについてSVMについて
SVMについて
 
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017
AI and Machine Learning Demystified by Carol Smith at Midwest UX 2017
 

Similar to DeepRemote: A Smart Remote Controller for Intuitive Control through Home Appliances Recognition by Deep Learning

DCE transmitters and receivers with Acio
DCE transmitters and receivers with AcioDCE transmitters and receivers with Acio
DCE transmitters and receivers with Aciokishorekumarsandan
 
(Slides) UbiREMOTE: Framework for Remotely Controlling Networked Appliances t...
(Slides) UbiREMOTE: Framework for Remotely Controlling Networked Appliances t...(Slides) UbiREMOTE: Framework for Remotely Controlling Networked Appliances t...
(Slides) UbiREMOTE: Framework for Remotely Controlling Networked Appliances t...Naoki Shibata
 
Semi-Real Time Inclinometer readings using Wireless Technologies
Semi-Real Time Inclinometer readings using Wireless TechnologiesSemi-Real Time Inclinometer readings using Wireless Technologies
Semi-Real Time Inclinometer readings using Wireless TechnologiesRekaNext Capital
 
Why do we all need to use CheckFox and what’s so Sigfox special about it?
Why do we all need to use CheckFox and what’s so Sigfox special about it?Why do we all need to use CheckFox and what’s so Sigfox special about it?
Why do we all need to use CheckFox and what’s so Sigfox special about it?Simple Hardware
 
Cairo 01 Six Sigma Measure
Cairo 01 Six Sigma MeasureCairo 01 Six Sigma Measure
Cairo 01 Six Sigma Measureahmad bassiouny
 
PID Control of Runaway Processes - Greg McMillan Deminar
PID Control of Runaway Processes - Greg McMillan DeminarPID Control of Runaway Processes - Greg McMillan Deminar
PID Control of Runaway Processes - Greg McMillan DeminarJim Cahill
 
Vocational training at DDK Delhi by SAKET RAI
Vocational training at DDK Delhi by SAKET RAIVocational training at DDK Delhi by SAKET RAI
Vocational training at DDK Delhi by SAKET RAISAKET RAI
 
DDK Delhi -Vocational training by Raisaket
DDK Delhi -Vocational training by RaisaketDDK Delhi -Vocational training by Raisaket
DDK Delhi -Vocational training by RaisaketSAKET RAI
 
Technical training 2009
Technical training 2009Technical training 2009
Technical training 2009Ilias Varsamis
 
Mpt presentation by manish abraham
Mpt presentation by manish abrahamMpt presentation by manish abraham
Mpt presentation by manish abrahamManish Abraham
 
Identifying Noise in DAQ
Identifying Noise in DAQIdentifying Noise in DAQ
Identifying Noise in DAQYokogawa1
 
Choosing the Right Sensors for Industrial Applications
Choosing the Right Sensors for Industrial ApplicationsChoosing the Right Sensors for Industrial Applications
Choosing the Right Sensors for Industrial ApplicationsDesign World
 
Identifying noise in daq webinar rp3 171027 (1)
Identifying noise in daq webinar rp3 171027 (1)Identifying noise in daq webinar rp3 171027 (1)
Identifying noise in daq webinar rp3 171027 (1)Kristin Allen
 
Hitachi set free features Compressors system Designing 919825024651
Hitachi set free features Compressors system Designing 919825024651Hitachi set free features Compressors system Designing 919825024651
Hitachi set free features Compressors system Designing 919825024651Sdsplit Split
 
SERENE 2014 Workshop: Paper "Using Instrumentation for Quality Assessment of ...
SERENE 2014 Workshop: Paper "Using Instrumentation for Quality Assessment of ...SERENE 2014 Workshop: Paper "Using Instrumentation for Quality Assessment of ...
SERENE 2014 Workshop: Paper "Using Instrumentation for Quality Assessment of ...SERENEWorkshop
 

Similar to DeepRemote: A Smart Remote Controller for Intuitive Control through Home Appliances Recognition by Deep Learning (20)

Article lior yosef
Article  lior yosefArticle  lior yosef
Article lior yosef
 
DCE transmitters and receivers with Acio
DCE transmitters and receivers with AcioDCE transmitters and receivers with Acio
DCE transmitters and receivers with Acio
 
(Slides) UbiREMOTE: Framework for Remotely Controlling Networked Appliances t...
(Slides) UbiREMOTE: Framework for Remotely Controlling Networked Appliances t...(Slides) UbiREMOTE: Framework for Remotely Controlling Networked Appliances t...
(Slides) UbiREMOTE: Framework for Remotely Controlling Networked Appliances t...
 
Semi-Real Time Inclinometer readings using Wireless Technologies
Semi-Real Time Inclinometer readings using Wireless TechnologiesSemi-Real Time Inclinometer readings using Wireless Technologies
Semi-Real Time Inclinometer readings using Wireless Technologies
 
Why do we all need to use CheckFox and what’s so Sigfox special about it?
Why do we all need to use CheckFox and what’s so Sigfox special about it?Why do we all need to use CheckFox and what’s so Sigfox special about it?
Why do we all need to use CheckFox and what’s so Sigfox special about it?
 
Qualification of high performance liquid chromatography
Qualification of high performance liquid chromatographyQualification of high performance liquid chromatography
Qualification of high performance liquid chromatography
 
Cairo 01 Six Sigma Measure
Cairo 01 Six Sigma MeasureCairo 01 Six Sigma Measure
Cairo 01 Six Sigma Measure
 
PID Control of Runaway Processes - Greg McMillan Deminar
PID Control of Runaway Processes - Greg McMillan DeminarPID Control of Runaway Processes - Greg McMillan Deminar
PID Control of Runaway Processes - Greg McMillan Deminar
 
★Data Logger Monitoring System★
★Data Logger Monitoring System★★Data Logger Monitoring System★
★Data Logger Monitoring System★
 
Thesis
ThesisThesis
Thesis
 
Vocational training at DDK Delhi by SAKET RAI
Vocational training at DDK Delhi by SAKET RAIVocational training at DDK Delhi by SAKET RAI
Vocational training at DDK Delhi by SAKET RAI
 
DDK Delhi -Vocational training by Raisaket
DDK Delhi -Vocational training by RaisaketDDK Delhi -Vocational training by Raisaket
DDK Delhi -Vocational training by Raisaket
 
Technical training 2009
Technical training 2009Technical training 2009
Technical training 2009
 
Mpt presentation by manish abraham
Mpt presentation by manish abrahamMpt presentation by manish abraham
Mpt presentation by manish abraham
 
Katalog fluke-building-thermal-imagers
Katalog fluke-building-thermal-imagersKatalog fluke-building-thermal-imagers
Katalog fluke-building-thermal-imagers
 
Identifying Noise in DAQ
Identifying Noise in DAQIdentifying Noise in DAQ
Identifying Noise in DAQ
 
Choosing the Right Sensors for Industrial Applications
Choosing the Right Sensors for Industrial ApplicationsChoosing the Right Sensors for Industrial Applications
Choosing the Right Sensors for Industrial Applications
 
Identifying noise in daq webinar rp3 171027 (1)
Identifying noise in daq webinar rp3 171027 (1)Identifying noise in daq webinar rp3 171027 (1)
Identifying noise in daq webinar rp3 171027 (1)
 
Hitachi set free features Compressors system Designing 919825024651
Hitachi set free features Compressors system Designing 919825024651Hitachi set free features Compressors system Designing 919825024651
Hitachi set free features Compressors system Designing 919825024651
 
SERENE 2014 Workshop: Paper "Using Instrumentation for Quality Assessment of ...
SERENE 2014 Workshop: Paper "Using Instrumentation for Quality Assessment of ...SERENE 2014 Workshop: Paper "Using Instrumentation for Quality Assessment of ...
SERENE 2014 Workshop: Paper "Using Instrumentation for Quality Assessment of ...
 

More from Yuta Takahashi

PikaBreak: 光で休憩を促してくれるスマートタイマー
PikaBreak: 光で休憩を促してくれるスマートタイマーPikaBreak: 光で休憩を促してくれるスマートタイマー
PikaBreak: 光で休憩を促してくれるスマートタイマーYuta Takahashi
 
ライフログを軸としたユビキタスコンピューティング技術
ライフログを軸としたユビキタスコンピューティング技術ライフログを軸としたユビキタスコンピューティング技術
ライフログを軸としたユビキタスコンピューティング技術Yuta Takahashi
 
生活リズムの類似性や周期性に基づく心身の健康状態の推定と予測
生活リズムの類似性や周期性に基づく心身の健康状態の推定と予測生活リズムの類似性や周期性に基づく心身の健康状態の推定と予測
生活リズムの類似性や周期性に基づく心身の健康状態の推定と予測Yuta Takahashi
 
観光中の内的状態推定に向けた観光客の無意識的しぐさの分析
観光中の内的状態推定に向けた観光客の無意識的しぐさの分析観光中の内的状態推定に向けた観光客の無意識的しぐさの分析
観光中の内的状態推定に向けた観光客の無意識的しぐさの分析Yuta Takahashi
 
生産性の低下抑止のためのウェアラブル機器による体調の推定と予測
生産性の低下抑止のためのウェアラブル機器による体調の推定と予測生産性の低下抑止のためのウェアラブル機器による体調の推定と予測
生産性の低下抑止のためのウェアラブル機器による体調の推定と予測Yuta Takahashi
 
歩行リハビリ支援のためのセンサ装着杖を 介した歩行動作認識手法の提案
歩行リハビリ支援のためのセンサ装着杖を 介した歩行動作認識手法の提案歩行リハビリ支援のためのセンサ装着杖を 介した歩行動作認識手法の提案
歩行リハビリ支援のためのセンサ装着杖を 介した歩行動作認識手法の提案Yuta Takahashi
 
TechUP Makersデモデイ - リハビリ支援のためのIoTデバイス『UbiCane』
TechUP Makersデモデイ - リハビリ支援のためのIoTデバイス『UbiCane』TechUP Makersデモデイ - リハビリ支援のためのIoTデバイス『UbiCane』
TechUP Makersデモデイ - リハビリ支援のためのIoTデバイス『UbiCane』Yuta Takahashi
 
労働生産性改善に向けたウェアラブル機器を用いた体調推定法の検討
労働生産性改善に向けたウェアラブル機器を用いた体調推定法の検討労働生産性改善に向けたウェアラブル機器を用いた体調推定法の検討
労働生産性改善に向けたウェアラブル機器を用いた体調推定法の検討Yuta Takahashi
 
茨城高専から九州と関西へ飛び出た話
茨城高専から九州と関西へ飛び出た話茨城高専から九州と関西へ飛び出た話
茨城高専から九州と関西へ飛び出た話Yuta Takahashi
 
HASCとWekaを使って行動認識
HASCとWekaを使って行動認識HASCとWekaを使って行動認識
HASCとWekaを使って行動認識Yuta Takahashi
 
Ubiquitous Tips - スキルウェンズデー
Ubiquitous Tips - スキルウェンズデーUbiquitous Tips - スキルウェンズデー
Ubiquitous Tips - スキルウェンズデーYuta Takahashi
 
交通事故データへの頻出パターンマイニングの適用
交通事故データへの頻出パターンマイニングの適用交通事故データへの頻出パターンマイニングの適用
交通事故データへの頻出パターンマイニングの適用Yuta Takahashi
 
筋電制御を支援するIoTプラットフォームと 把持戦略の構築
筋電制御を支援するIoTプラットフォームと把持戦略の構築筋電制御を支援するIoTプラットフォームと把持戦略の構築
筋電制御を支援するIoTプラットフォームと 把持戦略の構築Yuta Takahashi
 
Arch linuxを試したお話
Arch linuxを試したお話Arch linuxを試したお話
Arch linuxを試したお話Yuta Takahashi
 
超音波センサーを用いた4点杖の使用者のコンテキスト推定法の提案
超音波センサーを用いた4点杖の使用者のコンテキスト推定法の提案超音波センサーを用いた4点杖の使用者のコンテキスト推定法の提案
超音波センサーを用いた4点杖の使用者のコンテキスト推定法の提案Yuta Takahashi
 

More from Yuta Takahashi (16)

PikaBreak: 光で休憩を促してくれるスマートタイマー
PikaBreak: 光で休憩を促してくれるスマートタイマーPikaBreak: 光で休憩を促してくれるスマートタイマー
PikaBreak: 光で休憩を促してくれるスマートタイマー
 
ライフログを軸としたユビキタスコンピューティング技術
ライフログを軸としたユビキタスコンピューティング技術ライフログを軸としたユビキタスコンピューティング技術
ライフログを軸としたユビキタスコンピューティング技術
 
生活リズムの類似性や周期性に基づく心身の健康状態の推定と予測
生活リズムの類似性や周期性に基づく心身の健康状態の推定と予測生活リズムの類似性や周期性に基づく心身の健康状態の推定と予測
生活リズムの類似性や周期性に基づく心身の健康状態の推定と予測
 
観光中の内的状態推定に向けた観光客の無意識的しぐさの分析
観光中の内的状態推定に向けた観光客の無意識的しぐさの分析観光中の内的状態推定に向けた観光客の無意識的しぐさの分析
観光中の内的状態推定に向けた観光客の無意識的しぐさの分析
 
生産性の低下抑止のためのウェアラブル機器による体調の推定と予測
生産性の低下抑止のためのウェアラブル機器による体調の推定と予測生産性の低下抑止のためのウェアラブル機器による体調の推定と予測
生産性の低下抑止のためのウェアラブル機器による体調の推定と予測
 
歩行リハビリ支援のためのセンサ装着杖を 介した歩行動作認識手法の提案
歩行リハビリ支援のためのセンサ装着杖を 介した歩行動作認識手法の提案歩行リハビリ支援のためのセンサ装着杖を 介した歩行動作認識手法の提案
歩行リハビリ支援のためのセンサ装着杖を 介した歩行動作認識手法の提案
 
TechUP Makersデモデイ - リハビリ支援のためのIoTデバイス『UbiCane』
TechUP Makersデモデイ - リハビリ支援のためのIoTデバイス『UbiCane』TechUP Makersデモデイ - リハビリ支援のためのIoTデバイス『UbiCane』
TechUP Makersデモデイ - リハビリ支援のためのIoTデバイス『UbiCane』
 
労働生産性改善に向けたウェアラブル機器を用いた体調推定法の検討
労働生産性改善に向けたウェアラブル機器を用いた体調推定法の検討労働生産性改善に向けたウェアラブル機器を用いた体調推定法の検討
労働生産性改善に向けたウェアラブル機器を用いた体調推定法の検討
 
茨城高専から九州と関西へ飛び出た話
茨城高専から九州と関西へ飛び出た話茨城高専から九州と関西へ飛び出た話
茨城高専から九州と関西へ飛び出た話
 
HASCとWekaを使って行動認識
HASCとWekaを使って行動認識HASCとWekaを使って行動認識
HASCとWekaを使って行動認識
 
Ubiquitous Tips - スキルウェンズデー
Ubiquitous Tips - スキルウェンズデーUbiquitous Tips - スキルウェンズデー
Ubiquitous Tips - スキルウェンズデー
 
交通事故データへの頻出パターンマイニングの適用
交通事故データへの頻出パターンマイニングの適用交通事故データへの頻出パターンマイニングの適用
交通事故データへの頻出パターンマイニングの適用
 
筋電制御を支援するIoTプラットフォームと 把持戦略の構築
筋電制御を支援するIoTプラットフォームと把持戦略の構築筋電制御を支援するIoTプラットフォームと把持戦略の構築
筋電制御を支援するIoTプラットフォームと 把持戦略の構築
 
IoT Technologies
IoT TechnologiesIoT Technologies
IoT Technologies
 
Arch linuxを試したお話
Arch linuxを試したお話Arch linuxを試したお話
Arch linuxを試したお話
 
超音波センサーを用いた4点杖の使用者のコンテキスト推定法の提案
超音波センサーを用いた4点杖の使用者のコンテキスト推定法の提案超音波センサーを用いた4点杖の使用者のコンテキスト推定法の提案
超音波センサーを用いた4点杖の使用者のコンテキスト推定法の提案
 

Recently uploaded

Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Enterprise Knowledge
 
Vector Databases 101 - An introduction to the world of Vector Databases
Vector Databases 101 - An introduction to the world of Vector DatabasesVector Databases 101 - An introduction to the world of Vector Databases
Vector Databases 101 - An introduction to the world of Vector DatabasesZilliz
 
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Mark Simos
 
DevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsDevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsSergiu Bodiu
 
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024BookNet Canada
 
Commit 2024 - Secret Management made easy
Commit 2024 - Secret Management made easyCommit 2024 - Secret Management made easy
Commit 2024 - Secret Management made easyAlfredo García Lavilla
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebUiPathCommunity
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationRidwan Fadjar
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfAddepto
 
Training state-of-the-art general text embedding
Training state-of-the-art general text embeddingTraining state-of-the-art general text embedding
Training state-of-the-art general text embeddingZilliz
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Patryk Bandurski
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticscarlostorres15106
 
"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii SoldatenkoFwdays
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr LapshynFwdays
 
Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!Manik S Magar
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupFlorian Wilhelm
 
Search Engine Optimization SEO PDF for 2024.pdf
Search Engine Optimization SEO PDF for 2024.pdfSearch Engine Optimization SEO PDF for 2024.pdf
Search Engine Optimization SEO PDF for 2024.pdfRankYa
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationSlibray Presentation
 
Powerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time ClashPowerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time Clashcharlottematthew16
 

Recently uploaded (20)

Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024
 
Vector Databases 101 - An introduction to the world of Vector Databases
Vector Databases 101 - An introduction to the world of Vector DatabasesVector Databases 101 - An introduction to the world of Vector Databases
Vector Databases 101 - An introduction to the world of Vector Databases
 
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
 
DevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsDevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platforms
 
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
 
Commit 2024 - Secret Management made easy
Commit 2024 - Secret Management made easyCommit 2024 - Secret Management made easy
Commit 2024 - Secret Management made easy
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio Web
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 Presentation
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdf
 
Training state-of-the-art general text embedding
Training state-of-the-art general text embeddingTraining state-of-the-art general text embedding
Training state-of-the-art general text embedding
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
 
"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
 
Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project Setup
 
Search Engine Optimization SEO PDF for 2024.pdf
Search Engine Optimization SEO PDF for 2024.pdfSearch Engine Optimization SEO PDF for 2024.pdf
Search Engine Optimization SEO PDF for 2024.pdf
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck Presentation
 
Powerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time ClashPowerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time Clash
 

DeepRemote: A Smart Remote Controller for Intuitive Control through Home Appliances Recognition by Deep Learning

  • 1. DeepRemote: A Smart Remote Controller for Intuitive Control through Home Appliances Recognition by Deep Learning 〇Yuta Takahashi1, Naoki Shirakura1, Kenta Toyoshima1, Takuro Amako1, Ryota Isobe1, Jun Takamatsu1 and Keiichi Yasumoto1 1. Nara Institute of Science and Technology The Tenth International Conference on Mobile Computing and Ubiquitous Networking (ICMU2017) S7: Application, October 5, 2017
  • 2. Background ❖Increasing home appliances 2 ❖Many IR remote controllers ▪ Need to learn how to use ▪ High management costs Increase user’s burden
  • 3. Unification of remote controllers 3 Network Multiple IR controller Increase the number of appliances The interface becomes complicated Method for selecting home appliance is important! ✔ Management cost
  • 4. Home appliance selection 4 ❖With special attachments ❖Without special attachments ✔ Accurate selection Cost of devices/markers Selectable distance Voice [Pan 2010] Vision [Kong 2016] Intuitiveness ✔ Intuitiveness IR control Burden for wearing • IR LED [Neßelrath 2011] • IR transmitter [Tsukada 2004] • QR [Ullah 2012] Dedicated device is needed
  • 5. DeepRemote (Proposed) 5 ❖Home appliance selection Object recognition by deep learning → Intuitive and robust ❖Two units ▪ Control unit ▪ Deep learning unit ❖Network Home network (ROS)
  • 6. Control unit 6 ✓ Hand-held-type remote controller ❖Main processor ▪ Raspberry Pi3 (with Wi-Fi module) ❖Capturing a home appliance ▪ Front camera ❖Control interfaces ▪ Four buttons ▪ Gestures (right & left rotation)
  • 7. Deep learning unit 7 ✓ Image recognition ❖Main processor ▪ Laptop PC with Core i5 (Ubuntu) ▪ Distribute the calculation load of control unit ❖Recognition Model ▪ VGG16 [Simonyan 2014] TV Recognition result Deep learning unit Control unit
  • 8. Technique of training ❖Requirements of deep learning with zero-base 8 Huge dataset A lot of time ❖Fine tuning Trained model Arranging model Re-training
  • 10. Experimental environment 10 ✓Living room in smart-home facility ❖Five home appliances ▪ Fan (IR) ▪ Air conditioner (IR) ▪ TV (Network) ▪ Audio player (IR) ▪ Air purifier (Network) ❖Training data ▪ Took 20 images each appliances at P1, P2 and P3 ▪ 20 × 5 (appliances) × 3 (positions) = 300 images
  • 11. Example of training data 11 P1 P2 P3 Air purifier Air conditioner Audio player Fan TV
  • 12. Training ❖Model ▪ VGG16 trained ImageNet ❖Dataset ▪ Three hundred images ▪ Five classes 12 ❖Optimizer ▪ SDG ❖Loss function ▪ Categorical cross entropy
  • 13. Evaluations 1. Classification accuracy ▪ Captured 50 images of each appliance in each position ▪ 50 (images) × 5 (appliances) × 3 (positions) = 750 images 13 2. Response time ▪ Measured time since pushing button until return the result ▪ Evaluated at the same time as 1. 3. User test ▪ Verified control time of home appliances
  • 14. Appliance Precision [%] Recall [%] F-measure [%] P1 Air purifier 100.00 68.00 80.95 Audio player 83.33 90.00 86.54 TV 96.15 100.00 98.04 Air conditioner 57.47 100.00 72.99 Fan 86.96 40.00 54.80 Average 84.78 79.60 78.66 P2 Air purifier 81.63 80.00 80.81 Audio player 100.00 98.00 98.99 TV 92.59 100.00 96.15 Air conditioner 79.25 84.00 81.55 Fan 100.00 90.00 94.74 Average 90.69 90.40 90.45 P3 Air purifier 76.36 84.00 80.00 Audio player 100.00 58.00 73.42 TV 71.43 80.00 75.47 Air conditioner 53.17 84.00 65.12 Fan 100.00 62.00 76.54 Average 80.19 73.60 74.11 1. Classification accuracy of home appliances 14 P1:Effect of black door? P3: Too near?
  • 15. Evaluations 1. Classification accuracy ▪ Captured 50 images of each appliance in each position ▪ 50 (images) × 5 (appliances) × 3 (positions) = 750 images 15 2. Response time ▪ Measured time since pushing button until return the result ▪ Evaluated at the same time as 1. 3. User test ▪ Verified control time of home appliances
  • 16. 2. Response time 16 (n=750) Maximum time: 3.07 [sec] Minimum time: 1.72 [sec] Stable recognition about two seconds
  • 17. Evaluations 1. Classification accuracy ▪ Captured 50 images of each appliance in each position ▪ 50 (images) × 5 (appliances) × 3 (positions) = 750 images 17 2. Response time ▪ Measured time since pushing button until return the result ▪ Evaluated at the same time as 1. 3. User test ▪ Verified control time of home appliances
  • 18. 3. User test to assess control time (1/3) Experimental conditions Five participants Position: P1 Targets: Fan, Air conditioner, TV and Audio player 18 ❖Control time ❖Comparison Holding Power on DeepRemote Original vs
  • 19. 3. User test to assess control time (2/3) 19 F:86% F:98% F:77% F:55% High → ✔ about 5 seconds Lower than the 85% → over 10 seconds Accuracy Accuracy requires 85% over
  • 20. 3. User test to assess control time (3/3) 20 DeepRemote: Sum of each test > All Original: Sum of each test < All Changing time DeepRemote < Original Sum Sum 31.533.7 11.7 7.7 (- 2.1) (+ 3.0)
  • 21. Conclusions ❖DeepRemote ▪ Smart device for intuitively control the home appliances ▪ Deep learning method for home appliance selection ❖Results of evaluation ▪ 81.07% classification accuracy on average ▪ Average of response time is 1.97 seconds ▪ Time of power on an appliance takes 5 seconds ▪ 85% accuracy is required ▪ Time of changing target is lower than original remote controller 21
  • 23. Future works ❖Improving recognition accuracy ▪ Cropping object ❖Investigating energy consumption ❖Accuracy of distinguishing similar appliances ❖Experiment in long-term usage ▪ User teaches correct label when the system recognizes wrong 23
  • 24. Detail conditions of user test ❖Compare DeepRemote and original remote controller ▪ measure control time ▪ Remove “air conditioner” (no original controller) ▪ Five participants (males in the 20s) ▪ Perform three times in each measurement (one participant performs 30 times) ▪ Participant’s position: P1 24 1) Measure one appliance control time (start) → (power on an appliance) 2) Measure all appliances control time (start) → (power on the fan) →・ ・ ・ → (power on the audio player)
  • 25. Confusion matrix of P1 25 Classified Air purifier Air conditioner Audio player Fan TV True Air purifier 34 6 9 0 1 Air conditioner 0 50 0 0 0 Audio player 0 1 45 3 1 Fan 0 30 0 20 0 TV 0 0 0 0 50
  • 26. Confusion matrix of P2 26 Classified Air purifier Air conditioner Audio player Fan TV True Air purifier 40 6 0 0 4 Air conditioner 8 42 0 0 0 Audio player 1 0 49 0 0 Fan 0 5 0 45 0 TV 0 0 0 0 50
  • 27. Confusion matrix of P3 27 Classified Air purifier Air conditioner Audio player Fan TV True Air purifier 42 5 0 0 3 Air conditioner 7 42 0 0 1 Audio player 3 6 29 0 12 Fan 0 19 0 31 0 TV 3 7 0 0 40