Smart Crib Control System Based on Sentiment Analysis



Yüklə 0,61 Mb.
Pdf görüntüsü
səhifə4/7
tarix07.05.2023
ölçüsü0,61 Mb.
#109016
1   2   3   4   5   6   7
cameraready

while(!donedo 
/* select feature */ 
for each feature x in F do 
/*select the best feature x
+
*/ 
if E(result ∪ x
+
) = argmax[E(result ∪ x)] 
then
result
:= result ∪ x
+

F
:= F - x
+

end if 
/* stop when no features can be selected */ 
if no feature is selected 
then done := true; 
end if 
/* remove feature */ 
if (done = false) 
then 
repeat 
for each feature z in result do 
/* select the worst feature z
-
*/ 
if E(result - z
-
) = argmax[E(result - z)] 
then 
if E(result - z
-
) > E(result
then 
result
:= result - z
-

F
:= F + z
-

 
 
end if 
 
 
end if 
until no feature is removed 
end if 
end while 
Output: result; 


voting method. Once the model is trained, we use this model to 
predict crying states.
We give the pseudocode as follows. 
IV.
S
YSTEM 
I
MPLEMENTATION
While we already introduced the system architecture in 
Section III, we now provide further details on the components 
used in our system. 
Sensor: The sensors we use include pressure sensors, 
temperature sensors, sound sensors and humidity 
sensors. The data collected by the pressure sensor does 
not need to be processed, and the data obtained by other 
sensors all need to be processed to determine the next 
action taken by the system. These sensors are all native 
sensors of the Raspberry Pi, which means that they are 
all compatible with the Raspberry Pi. 
Raspberry Pi: We use the Raspberry Pi as user 
interface to control the hardware. This interface allows 
the user to modify some settings, such as the frequency 
and amplitude of crib sway, music played, etc. The 
graphical user interface is displayed on an external LCD 
display connected to the Raspberry Pi. After connecting 
with the pin of the Raspberry Pi expansion board, the 
user can set up the device. The GUI is shown in Fig. 5. 
The data processing algorithms, including speech pre-
processing, SVM algorithms, etc, are implemented in 
Python language.  
Server: The servlets on the server are implemented in 
Java programming language. There are several servlets 
to response particular requests. All servlets are running 
on Tomcat which is equipped on an ALiYun server. 
Tomcat is a container of servlets that can store all the 
servlets and enables them to run on the server. The data 
sending to or receiving from servlets are in JSON 
format, which is a uniform communication format. 
Mobile Terminal: So far, we have only developed an 
App for Android phones as Android is the most 
commonly used operational system for smartphones. 
Java programming language using the Android 
Software Development Kit (SDK) has been used for the 
development and implementation of the App. The main 
functions on this App is showing baby
’s data in graphs 
and charts and uploading baby
’s videos as well as 
photos. The app also allows parents to set the hardware, 
like playing particular music, setting buzzer
’s volume, 
etc. With the help of this App, parents can see the 
growth of their baby directly and know the next move 
of their baby so that they can make some changes to 
meet the baby
’s requirement.  
V.
E
XPERIMENTAL 
R
ESULTS
The experiment were performed on a computer with an 
Intel(R) Core(TM) i7-7700HQ processor and 8GB of main 
memory, running Microsoft Windows 10 Professional.
All crying data we used for the experiment has been 
extracted from videos of crying babies that have been shared on 
the YouTube platform. Our dataset includes five male and six 
female babies of different ethnicity, i.e., three Asian babies, five 
Caucasian babies, and three Black babies. Their sentiment is 
labeled according to the title of the video and assessed by a 
professional nurse. The non-crying data also comes from the 
Internet, including silence, noise, laughter, chicken roar
barking, meows, footsteps, etc. 
In the remainder of this section we show the results of the 
algorithm described in Section Ⅲ, including the result of the 
crying preprocess and the result of classification.
对话框标题
Rotation amplitude adjustment

WIFI settings
Music settings
Current connection: none
Current music: none
Volume
Edit
Eidt
Figure 5. User Interface of the Raspberry Pi 
Input: f is the feature vector extracted; 
C[]
are the candidate classes; 
S()
is the trained SVM classifier; 
V[]
are the number of votes of the classes; 

Yüklə 0,61 Mb.

Dostları ilə paylaş:
1   2   3   4   5   6   7




Verilənlər bazası müəlliflik hüququ ilə müdafiə olunur ©azkurs.org 2024
rəhbərliyinə müraciət

gir | qeydiyyatdan keç
    Ana səhifə


yükləyin