eTudlo: A Mobile Learning Tool for Pre-School in

Object Identification using OpenCV

Geralyn P. Masuhay

Diane Rose G. Calzada

Allvel P. Gorne

A Capstone Project

Submitted in Partial Fulfillment

Of the Requirements for the Degree

BACHELOR OF SCIENCE IN INFORMATION TECHNOLOGY

Department of Information Technology Education

ACLC College of Butuan
Butuan City, 8600 Philippines

AUGUST 2019
 CHAPTER 1
INTRODUCTION

1.1 Introduction

Mobile learning is an exciting and important movement. In our ever-changing

marketplace, today pre-schools students are being educated. Largely, for jobs

that have not yet been invented. Technology is being created and updated at a

frenetic pace, and growing more pervasive and useful with each stride.

According to (Franklin Delano Rosevelt, 2017) we cannot always build the

future for our youth, but we can build our youth for the future. As schools face

dwindling resources and higher stakes than ever, does mobile technology hold

the key in improving the educational system and re-engaging pre-school

students. Mobile Technology offers plethora features and benefits that enables

it to break the education system wide open, engaging pre-schools students in

ways and making educational experiences more meaningful, if schools can

effectively utilize structured, integrated approaches for implementation of this

new technology.

The curriculum of the preschool is the National Kindergarten Curriculum

Guide based on the Republic Act 8980 known as the Early Childhood Care

Development (ECCD) Act this is an act promulgating a comprehensive national

system for early childhood care and development, providing funds therefore

and for the other purposes. As children grow into early childhood, their world

will begin to open up. They will become more independent and begin to focus

more on adults and children outside of the family. They will want to explore and

ask about the things around them even more. Preschool curricula take into
account the length of the preschool day. Many preschools run for only a few

hours a day, while others (especially in public school settings) run for much

longer days. Some even run longer than a typical school day to cover all the

parent's work hours (Amanda Rock, 2019)

E‐learning as the appropriate application of the Internet to support the

delivery of learning, skills, and knowledge in a holistic approach not limited to

any particular courses, technologies, or infrastructures. It considers e‐learning

core products of content, technology, and services. The emphasis is upon the

importance of integrated learning paths that vary according to the student, the

subject material, the level of competence, and corporate or student

preferences. Successful implementation of e‐learning requires the same

management commitment as other mission‐critical organization‐wide

initiatives. Most of all e‐learning needs to be compelling to the audience it

targets, offering the learner a resource that is seen as appealing, valuable and

productive to their goals and aspirations (Paul Henry, 2001)

The goal of object detection is to detect all instances of objects from a known

class, such as people, cars or faces in an image. Typically, only a small number

of instances of the object are present in the image, but there is a very large

number of possible locations and scales at which they can occur and that need

to somehow be explored. Add the importance of audio in retention rate in

learning.This study aimed to provide a better way to teach the young ones to

identify the basic objects that they need to recognize and in this way the

learning of the children will as easy as it can be.

1.2 Project Context

Nowadays preschool education is considered crucial for a child’s

development. Recent researches support the view that the role of

kindergarten in children’s progress is very important, as this age is

sensitive for their future academic and social life. Based on the survey

the proponent conducted all of the teacher said that the students in

preschool are very active in object identifying. The teacher also uses

power-points or television if the teacher demonstrate the students

Smartphone the device suitable for visual and auditory as agents in

learning. The use of mobile learning is recognized as a tool that can

foster the knowledge and the experiences for this age and the support

of specific areas according to the educational perspective. In this paper

we try to give a brief overview of the most representative studies of the

last decade (2005-2015), which focus on the skills that are explored in

kindergarten and are supported by the mobile applications. The

effectiveness of mobile learning in special preschool education is also

explored. People are using a smartphone for their everyday lives. An

android phone that has many applications or games that installed to it. It

is very possible to also use technology in learning especially on children

age 3-5 years old in preschool, it will catch their attention quickly and the

teacher can use that as the perfect time on teaching the things that they

have to learn and the things they have to familiarize. There are few steps

to use the application, the user will simply face the camera to the object
 around the classroom or study room and it will automatically recognize

the object it detected and the application will read the label of the object

one at a time. In this application, the researchers will use OpenCV as

the main library for identifying the object and Kulang pa

1.3 Purpose and Description

The purpose of the study is to provide a tutorial application for the

children at the age of 3-5 years old who are on the stage of learning, by this

they can easily familiarized the object they encounter inside their study room or

classroom. The application can assist the teachers in educating the children in

a current way.

1.4 The objective of the study

General Objective

The main objective of the study is to develop an object detection

identifier using OpenCV with auditory as a tutorial application for preschool

students.

Specific Objective

The following are the specific objective of the study:

 To develop a tutorial application that would identify the objects

around the classroom based on the theme or the quarter of the class

based in the ECCD Curriculum.

  To develop an application for tutoring children the basic things they

need to recognize.

 To evaluate the develop system in terms of:

o Functionality;

o Reliability;

o Efficiency and;

o Usability.

1.5 Scope and Limitation of the Study

1.5.1 Scope

The following are the scope of the system:

 The application can detect multiple objects within an image and read the

label of the objects one at a time.

 Applicable to preschool students and those who are willing to learn.

 The application will read the label of the object it detected.

1.5.2 Limitation

The following are the limitation of the system:

 There are thirty (30) different classes of objects to be detected.

 The application is not available online.

 If the object is not detected clearly, the application will not

recognize it.

 The application can only read the object it detected.

1.6 Definition of Terms

The following terms are conceptually and operationally defined for a better

understanding of the study.

 OpenCV - (Open source computer vision) is a library of

programming functions mainly aimed at real-time computer vision.

Originally developed by Intel, it was later supported by Willow

Garage then Itseez (which was later acquired by Intel)

 Real-Time- the actual time during which a process or event occurs.

1. https://www.researchgate.net/publication/309446775_Mobile_Learning_For_Presc

hool_Education#targetText=Mobile%20Learning%20For%20Preschool%20Education

&targetText=Nowadays%20preschool%20education%20is%20considered,future%20

academic%20and%20social%20life.

2. https://www.verywellfamily.com/all-about-preschool-curriculum-2764977
 3. https://www.cdc.gov/ncbddd/childdevelopment/positiveparenting/preschoolers.ht

ml

CHAPTER 2

REVIEW OF RELATED LITERATURE AND SYSTEM

This chapter includes the ideas, finished thesis, generalization or

conclusions, methodologies, and others. Those that were included in this

chapter helps in familiarizing information that was relevant and similar to the

present study.

2.1 Related Systems

2.1.1 Human Skin Color Detection Using Bayesian Rough Decision Tree

(2018)

Detecting human skin color is a very difficult function and these are the

two reasons, mainly changeable illumination conditions and diverse races of

people according to Ayad R. Abbas, Ayat O. Farooq (2018). Some previous

researchers in this field tried to resolve those problems by using thresholds that

relied on specific values of skin tones. Although, it is a speedy and an easy

implementation, it does not provide sufficient information for recognizing all skin

tones of humans.

This paper proposes Bayesian Rough Decision Tree (BRDT) classifier

to improve the accuracy of human skin detection. Three experiments have been

conducted using (RGB) dataset collected from University of California, Irvine

(UCI) machine learning repository, RGB (Red, Green, Blue), HSV (Hue,

Saturation, Value) and YCbCr (Luminance, Chrominance).

This said study was similar to the current study in the way that they

are detecting human skin color. The system of the said study also shows that

the proposed system can achieve preferable accuracy in skin detection 98%,

97% and 97% using RGB dataset, HSV dataset and YCbCr dataset

respectively.

2.1.2 Object Detection Using Image Processing (2016)

The Image processing is a method to convert an image into digital form

and perform some operations on it, in order to get an enhanced image or to

extract some useful information from it as indicated by Fares Jalled, Ilia

Voronkov (2016). It is a type of signal dispensation in which input is image, like

video frame or photograph and output may be image or characteristics

associated with that image. Usually Image Processing system includes treating

images as two-dimensional signals while applying already set signal processing

methods to them. Object detection determines the presence of an object and/or

its scope, and locations in the image.

Object Detection Using Image Processing and the current study of the

proponents are just the same for it talks about Object Detection. The system of

the said study can also analyze and manipulating the image which includes

data compression and image enhancement and spotting patterns that are not

to human eyes like satellite photographs.

2.1.3 A real-time system for object detection and location reminding with

RGB-D camera (2014)

The study presents a helpful application with a real-time detection

system that can automatically capture the last scene where the user-defined

important objects appear. The introduced method uses RGB-D information as

input and has high detection rate in messy indoor environments. Additionally,

we build a user-friendly using flow on object online learning and detection which

may be suitable for future technologies of wearable devices. I-Kuei Chen,

Chung-Yu Chi, Szu-Lu Hsu, Liang-Gee Chen, “An integrated system for object

tracking detection and online learning with real-time RGB-D video”, Acoustics

Speech and Signal Processing (ICASSP) 2014 IEEE International Conference

on, pp. 6558-6562, 2014.

 While the study of the proponent indicates similarities particularly in real-

time detection that also detects object and process or identities through the

method RGB-D information. Both studies are focusing on detecting particular

object that has a high percentage of having similar characteristics with accurate

reading and processing series of data to make or give an automatic description

of the particular object.

2.1.4 Autonomous Learning of Object Models on a Mobile Robot (2017)

The study presents and evaluate a system, which allows a mobile robot

to autonomously detect, model, and re-recognize objects in everyday

environments according to Thomas Faulhammer, Rares Ambrus, Chris

Burbridge, Michael Zillich, John Folkesson, Nick Hawes, Patric Jensfelt,

Markus Vincze (2017) While other systems have demonstrated one of these

elements, to our knowledge, they present the first system, which is capable of

doing all of these things, all without human interaction, in normal indoor scenes.

The system detects objects to learn by modeling the static part of the

environment and extracting dynamic elements. It then creates and executes a

view plan around a dynamic element to gather additional views for learning.

This said study was similar to the current study in the way that it also

talks about detecting and recognizing objects. The system of the said study can

also evaluated on publicly available datasets as well as on data collected by

the robot in both controlled and uncontrolled scenarios.

2.1.5 Location-Based Augmented Reality for Mobile Learning: Algorithm,

System, and Implementation (2015)

 AR technology can be considered as mainly consisting of two aspects:

identification of real-world object and display of computer-generated digital

contents related the identified real-world object as indicated by Tan, Qing;

Chang, William; Kinshuk (2015). The technical challenge of mobile AR is to

identify the real-world object that mobile device's camera aim at. In this paper,

they present a location-based object identification algorithm that has been used

to identify learning objects in the 5R adaptive location-based mobile learning

setting.

The said system was similar to the current study based on the

technicalities presented. This study also provide some background of the

algorithm, discuss issues in using the algorithm, and present the algorithm

empowered mobile learning system and its implementation.

2.1.6 DeepMon: Mobile GPU-based Deep Learning Framework for

Continuous Vision Applications (2014)

The rapid emergence of head-mounted devices such as the Microsoft

Holo-lens enables a wide variety of continuous vision applications. According

to Loc N. Huynh, Youngki Lee, Rajesh Krishna Balan such applications often

adopt deep-learning algorithms such as CNN and RNN to extract rich

contextual information from the first-person-view video streams. Despite the

high accuracy, use of deep learning algorithms in mobile devices raises critical

challenges. The study proposes DeepMon, a mobile deep learning inference

system to run a variety of deep learning inferences purely on a mobile device

in a fast and energy-efficient manner. For this, we designed a suite of

optimization techniques to efficiently offload convolutional layers to mobile

GPUs and accelerate the processing; note that the convolutional layers are the

common performance bottleneck of many deep learning models

This said study was similar to the current study in the way that it also

talks about Mobile Learning. The system of the said study can also classify an

image over the VGG-VeryDeep-16 deep learning model in 644ms on Samsung

Galaxy S7, taking an important step towards continuous vision without imposing

any privacy concerns or networking cost.

2.1.7 Real Time Object Scanning Using a Mobile Phone and Cloud-based

Visual Search Engine (2017)

Computer vision and human-powered services can provide blind people

access to visual information in the world around them, but their efficacy is

dependent on high-quality photo inputs. Blind people often have difficulty

capturing the information necessary for these applications to work because they

cannot see what they are taking a picture of. According to Yu Zhong, Pierre J.

Garrigues, Jeffrey P. Bigham (2017) The study presents Scan Search, a mobile

application that offers a new way for blind people to take high-quality photos to

support recognition tasks. To support real-time scanning of objects, they

developed a key frame extraction algorithm that automatically retrieves high-

quality frames from continuous camera video stream of mobile phones.

Those key frames are streamed to a cloud-based recognition engine that

identifies the most significant object inside the picture. This way, blind users

can scan for objects of interest and hear potential results in real time. The study

also present a study exploring the tradeoffs in how many photos are sent, and

conduct a user study with 8 blind participants that compare Scan Search with
a standard photo-snapping interface. Our results show that Scan Search allows

users to capture objects of interest more efficiently and is preferred by users to

the standard interface.

2.1.8 Augmented reality, the future of contextual mobile learning (2016)

This study aims to show the relevance of augmented reality (AR) in

mobile learning for the 21st century. With AR, any real-world environment can

be augmented by providing users with accurate digital overlays. As indicated

by Roopesh Kevin Sungkur, Akshay Panchoo, Nitisha Kirtee Bhoyroo (2016)

AR is a promising technology that has the potential to encourage learners to

explore learning materials from a totally new perspective. Besides, the

advancements made in information technology further broaden the scope for

educational AR applications. Furthermore, the proliferation of wireless mobile

devices such as smartphones and tablets is also introducing AR into the mobile

domain.

The study gives an insight of the different potential fields of application

of AR and eventually proposes an AR application that will give a completely

different learning experience for learners. This AR mobile application will not

only provide learners with supplementary information but will also assist

lecturers in their teaching process. There are certain concepts in computer

science at the tertiary level that are at times difficult for learners to understand

using the traditional classroom approach. Through this AR application

developed, the learners are able to see what is happening and experience a

different form of learning where the focus is more on “learning by doing” and on
the ability of visualizing the complete set of steps involved for a particular

operation. Finally what is proposed is a generic framework/process for the

development of AR applications for learning purposes.

2.1.9 DeepCham: Collaborative Edge-Mediated Adaptive Deep Learning

for Mobile Object Recognition (2016)

Deep learning techniques achieve state-of-the-art performance on many

computer vision related tasks. According to Dawei Li, Theodoros Salonidis,

Nirmit V. Desai, Mooi Choo Chuah (2016) the study shows that recognition

accuracy degrades when used in daily mobile scenarios due to context

variations caused by different locations, time of a day, etc. To solve this

problem, we present DeepCham - the first adaptive mobile object recognition

framework that allows deep learning techniques to be used successfully in

mobile environments. Specifically, DeepCham is mediated by an edge master

server which coordinates with participating mobile users to collaboratively train

a domain-aware adaptation model which can yield much better object

recognition accuracy when used together with a domain-constrained deep

model. DeepCham generates high-quality domain-aware training instances

for adaptation from in-situ mobile photos using two major steps: (i) a distributed

algorithm which identifies qualifying images stored in each mobile device for

training, (ii) a user labeling process for recognizable objects identified from

qualifying images using suggestions automatically generated by a generic deep

model.
 Using a newly collected dataset with smartphone images collected from

different locations, time of a day, and device types, we show that DeepCham

improves the object recognition accuracy by 150% when compared to that

achieved merely using a generic deep model. In addition, we investigated how

major design factors affect the performance of DeepCham. Finally, we

demonstrate the feasibility of DeepCham using an implemented prototype.

2.2 Related Literature

2.2.1 Faster R-CNN: Towards Real-Time Object Detection with Region

Proposal Networks (2015)

State-of-the-art object detection networks depend on region proposal

algorithms to hypothesize object locations. Advances like SPPnet and Fast R-

CNN have reduced the running time of these detection networks, exposing

region proposal computation as a bottleneck. In this work, we introduce a

Region Proposal Network (RPN) that shares full-image convolutional features

with the detection network, thus enabling nearly cost-free region proposals. An

RPN is a fully convolutional network that simultaneously predicts object bounds

and objectness scores at each position. RPNs are trained end to end to

generate high-quality region proposals, which are used by Fast R-CNN for

detection. With a simple alternating optimization, RPN and Fast R-CNN can be

trained to share convolutional features. For the very deep VGG-16 model, our

detection system has a frame rate of 5fps (including all steps) on a GPU, while

achieving state-of-the-art object detection accuracy on PASCAL VOC 2007

(73.2% mAP) and 2012 (70.4% mAP) using 300 proposals per image. Code is

available at https://github.com/ShaoqingRen/faster_rcnn.
 This study is related to the proponents study due to the fact that it also

uses real-time object detection and it is somehow similar to the proponents’

application. The difference is that it has an added feature that is the region

proposal networks.

2.2.2 Research on Daily Objects Based on Deep Neural Network (2018)

With the rapid development of deep learning, great breakthroughs have

been made in the field of object detection. In this article, the deep learning

algorithm is applied to the detection of daily objects, and some progress has

been made in this direction. Compared with traditional object detection

methods, the daily objects detection method based on deep learning is faster

and more accurate. The main research work of this article: 1. Collect a small

data set of daily objects; 2. In the Tensor flow framework to build different

models of object detection, and use this data set training model; 3. the training

process and effect of the model are improved by fine-tuning the model

parameters. Ding (2018). Research on Daily Object Detection Based on Deep

Neural Network.

The study of Ding uses machine learning to identify the object

accurately. Relating to the proponents study in using tensor flow as the

framework for machine learning and has the capability to enhance the clarity of

detecting a particular object in real-time. With the new research about deep

neural network, advance algorithm may also adapted by the proponents study

to enhance the capability of detecting and comparing objects to find a particular

object that is requisite and required for the user to understand the information

given by the system that is detected real-time.

2.2.3 You Only Look Once: Unified, Real-Time Object Detection (2015)

You Only Look Once: Unified, Real-Time Object Detection (2015) is a

new approach to object detection. According to Joseph Redmon, Santosh

Divvala, Ross Girshick, Ali Farhadi (2015) YOLO Prior work on object detection

repurposes classifiers to perform detection. Instead, we frame object detection

as a regression problem to spatially separated bounding boxes and associated

class probabilities. A single neural network predicts bounding boxes and class

probabilities directly from full images in one evaluation. Since the whole

detection pipeline is a single network, it can be optimized end-to-end directly

on detection performance.

Our unified architecture is extremely fast. Our base YOLO model

processes images in real-time at 45 frames per second. A smaller version of

the network, Fast YOLO, processes an astounding 155 frames per second

while still achieving double the mAP of other real-time detectors.

This study has a similarity to the current study in the way that it talks

about processes images in real-time. The system of the said study also makes

more localization errors but is far less likely to predict false detections where

nothing exists. It also learns very general representations of objects.

2.2.4 The effectiveness of a meaningful learning‐based evaluation model

for context‐aware mobile learning (2014)

In recent years, context‐aware mobile learning (CAML) has been widely

applied to various fields and has become a popular issue in educational

research. Despite the tremendous potential of CAML and its growing

significance, continued evaluations and refinements under the advice of field

experts and instructors are crucial to ensure its validity, value and sustainability.

As indicated by Yueh-Min Huang, Po-Sheng Chiu (2014), an evaluation model

for CAML is developed based on meaningful learning theory using the analytic

hierarchy process (AHP). To verify the effectiveness of the model, three

different CAML learning activities are tested, and some experts are invited to

evaluate and compare them. As a result, the strengths and weaknesses of each

learning activity are obtained. With the aid of the evaluation model proposed in

this work, CAML developers can progressively enhance the value of learning

activities, thus improving this new approach to learning.

This study has a similarity to the current study in the way that it talks

about mobile learning fields. The system of the said study also makes more

popular issue in educational research. It can also verify the effectiveness of the

model.

2.2.5 Object recognition and detection with deep learning for autonomous

driving applications (2017)

 Autonomous driving requires reliable and accurate detection and

recognition of surrounding objects in real drivable environments. Although

different object detection algorithms have been proposed, not all are robust

enough to detect and recognize occluded or truncated objects. According to

Aysegul Ucar, Yakup Demir, Cuneyt Guzelis (2017) the study propose a novel

hybrid Local Multiple system (LM-CNN-SVM) based on Convolutional Neural

Networks (CNNs) and Support Vector Machines (SVMs) due to their powerful

feature extraction capability and robust classification property, respectively.

The proposed system has the whole image into local regions and employ

multiple CNNs to learn local object features. The system select discriminative

features by using Principal Component Analysis. We then import into multiple

SVMs applying both empirical and structural risk minimization instead of using

a direct CNN to increase the generalization ability of the classifier system.

Finally, we fuse SVM outputs. In addition, the system pre-trained AlexNet and

a new CNN architecture. We carry out object recognition and pedestrian

detection experiments on the Caltech-101 and Caltech Pedestrian datasets.

Comparisons to the best state-of-the-art methods show that the proposed

system achieved better results.

2.2.6 A Review of Detection and Tracking of Object from

Image and Video Sequences (2017)

Object detection and tracking is one of the critical areas of research due

to routine change in motion of object and variation in scene size, occlusions,

appearance variations, and ego-motion and illumination changes.

Specifically, feature selection is the vital role in object tracking. As indicated

by Mukesh Tiwari, Dr. Rakesh Singhai (2017) It is related to many real time

applications like vehicle perception, video surveillance and so on. In order to

overcome the issue of detection, tracking related to object movement and

appearance. Most of the algorithm focuses on the tracking algorithm to

smoothen the video sequence. On the other hand, few methods use the prior

available information about object shape, color, texture and so on. Tracking

algorithm which combines above stated parameters of objects is discussed

and analyzed in this research. The goal of this paper is to analyze and review

the previous approach towards object tracking and detection using video

sequences through different phases.

This study has a similarity to the current study in the way that it talks

about tracking and object detection. The system of the said study also

identify the gap and suggest a new approach to improve the tracking of

object over video frame.

2.2.7 A Critical Review of Mobile Learning Integration in Formal

educational contexts (2018)

The use of digital technology in the learning process and teaching practices

in formal teaching is highly dependent on the ability of teachers of introducing

it without jeopardizing the richness of the classroom environment, namely the

attention that students need to follow the flow of argumentation and to

guarantee the quality of the inquiring. According to Luis Francisco Mendes

Gabriel Pedro, Claudia Marina Monica de Oliveira Barbosa and Carlos Manuel
das Neves (2018) Although several studies value the importance of

technologies in our media-enriched world and the learn anytime and anywhere

motto associated with mobile learning, the argue that the classroom dynamics

are becoming more and more at risk with the addictive dimension brought about

by the ubiquitous presence of digital devices and social media in students’ lives.

The study of the said system make a critical review of the literature

related to mobile learning because there is still a need of more extensive

research on the interference of technology in the classroom, especially on how

multitasking affects the teacher role in-class as a media orchestrator and

learning facilitator. Finally, the system discuss the use of technology in the

formal classroom environment, mainly to stimulate a much-needed discussion

about the bright-not-so-bright impacts of technology in the teaching and

learning process.

2.2.8 A Literature Survey: Neural Networks for Object Detection (2018)

Humans have a great capability to distinguish objects by their vision. But,

for machines object detection is an issue. Thus, Neural Networks have been

introduced in the field of computer science. As indicated by Aishwarya Sarkale,

Kaiwant Shah, Anandji Chaudhary, Tatwadarshi (2018) Neural Networks are

also called as ‘Artificial Neural Networks’ [13]. Artificial Neural Networks are

computational models of the brain which helps in object detection and

recognition. This paper describes and demonstrates the different types of

Neural Networks such as ANN, KNN, FASTER R-CNN, 3D-CNN, RNN etc. with

their accuracies. From the study of various research papers, the accuracies of
different Neural Networks are discussed and compared and it can be concluded

that in the given test cases, the ANN gives the best accuracy for the object

detection.

This study has a similarity to the current study in the way that it talks

about object detection. The system of the said study have an Artificial Neural

Networks, it is a type of artificial intelligence that attempts to simulate the way

a human brain works. Rather than using a digital model, in which all

computations manipulate zeros and ones, a Neural Network works by creating

connections between processing elements, the computer equivalent of

neurons.

2.2.9 Mobile Object Detection using TensorFlow Lite and Transfer

Learning (2018)

With the advancement in deep learning in the past few years, the

system are able to create complex machine learning models for detecting

objects in images, regardless of the characteristics of the objects to be

detected. According to Oscar Alsing (2018) This development has enabled

engineers to replace existing heuristics based systems in favour of machine

learning models with superior performance. In this report, we evaluate the

viability of using deep learning models for object detection in real-time video

feeds on mobile devices in terms of object detection performance and

inference delay as either an end-to-end system or feature extractor for existing

algorithms.

This study has a similarity to the current study in the way that it talks

about Mobile Object Detection. The system of the said study shows a

significant increase in object detection performance in comparison to existing

algorithms with the use of transfer learning on neural networks adapted for

mobile use.

2.2.10 A Survey of Modern Object Detection Literature using Deep

Learning (2018)

Object detection is the identification of an object in the image along with

its localization and classification. It has wide spread applications and is a critical

component for vision based software systems. As indicated by Karanbir Chahal

and Kuntal Dey (2018) the system seeks to perform a rigorous survey of

modern object detection algorithms that use deep learning. As part of the

survey, the topics explored include various algorithms, quality metrics,

speed/size tradeoffs and training methodologies. This paper focuses on the two

types of object detection algorithms- the SSD class of single step detectors and

the Faster R-CNN class of two step detectors. Ultimately, a rigorous review of

the strengths and weaknesses of each detector leads us to the present state of

the art.
 This study has a similarity to the current study in the way that it talks

Modern Object Detection. The system of the said study have a techniques to

construct detectors that are portable and fast on low powered devices are also

addressed by exploring new lightweight convolutional base architectures.

CHAPTER 3

DESIGN AND METHODOLOGY

This chapter presents the discussion of the design and methodology of

the proposed system.

3.1 The Project Concepts

Figure 1. Conceptual Framework of the System

Figure 1 illustrate the physical movement of the application. It specifies

the data for user input, the process of the system, the output and also the

outcome of the system.

3.2 Systems Analysis and Design

3.2.1 User Requirements

Table 1. User Requirements

Name of User Main Module Sub Modules

Project
 eTudlo: A

Mobile

Learning Object

Tool for Pre- User Object Identifier Information

School in

Object

Identification

using Open

CV

List of Objects Object information

chair  Description

 Attribute

3.2.2 Functional Requirements

3.2.2.1 Operational Environment

 Figure 2. Operational Environment

The figure 2 shows how the users access the application through using a mobile

phone.

3.2.2.2 System Interface

The following are the interfaces that the proposed application

used.

 The app works in MySQL as its database server

 OpenCV for real-time identifier

3.2.2.3 Communication Interface

The following are the communication interfaces used.

  Python – Python was a popular programming language

having simple, expressive and elegant syntax that helps in

creating an application. The proponents would use this

framework to have a clean and simple routing.

 Open CV - (Open Source Computer Vision) is an open

source computer vision and machine learning software

library.

3.2.2.4 Software Interface

The following are the software interfaces being used.

 MySQL- used as the database server.

 Sublime 3.1 – use a text/code editor.

3.2.2.5 Hardware Interface and Architecture Design

Table 2. Hardware Interface

Hardware Interface Function

Laptop Use for the application development

and logical operations.

Smartphone For testing the application.

Table 2 shows the hardware interface used by the proponents in

developing the system

3.2.2.6 Function/User Security Matrix

 The following symbols represent the level of access by each of the

user groups:

R Read

S Scan

Table 3. Function/User Security Matrix

Function Actor Access Level

Identify Object USER S

Object Information USER R

3.2.2.7 User Group and System Access Summary

Table 4. User Groups and System Access Summary

User Group System Access

Are authorized to all function and the

User
capabilities of the application.

3.2.3 Non-Functional Requirements

3.2.3.1 Reliability

The application is able to identify an object.

3.2.3.2 Operability

Only the user can operate the application.

3.2.3.3 Maintainability

It will use a Smartphones to work the application.

3.2.3.4 Scalability
 The application has only 1 user types.

3.2.3.5 Availability

The application works in Smartphones.

3.2.3.6 Delivery

 The application will be delivering to the Pre-school

students. The proponents are the one responsible in

installing the application in smartphones to make the it

ready to use and proponents would be the one responsible

for conducting training to the users.

3.2.4 System Requirements

3.2.4.1 Hardware Requirements

Hardware Specification

Processor 3.50 GHz

RAM 4GB RAM

Hard Disk Drive 500 MB HDD

Video Card 512mb or higher

Android Phone 2 GB

3.2.4.1 Software Requirements

 Software Specification

Android Studio Version 3.2.1

Xampp Version 7.1.24

Windows OS 7, 8 or 10

Android 6.0 Android Phone

Sublime Version 3.1.1

3.3 Development Model of the System

Figure 3. The phase of the System Development Life Cycle RAD (Rapid

Application Development)

The proponents used the System Development Life Cycle

(SDLC) method and RAD (Rapid Application Development) the functional

modules are developed in parallel as prototypes and are integrated to make the

complete product for faster product delivery.

3.4 Development Approach

Figure 4. Top-Down Approach

The proponents decided to use the top-down approach which was

essential in breaking down the development of application. The proponents first

do the planning of the application after that is the requirements specification

and analysis which is important in order to formulate the expected outcome of

the application, then the designing and program coding is the phase which is

the application’s front-end and back-end will be develop, after the designing

and program coding is the testing which the application will be tested if it meets

the expected outcome, maintenance and modification which includes the

maintenance of the application.

3.4.1 Requirement Analysis

During the requirement gathering and analysis phase, the proponents

gather enough data from the respondents and make notable information that is

discussed to the respondent for verification purpose. The analysis that the

proponent conducted is on schedule.

3.4.2 Design

3.4.2.1 DFD Level 0

Object

USER eTudlo: A Mobile Learning Tool for

Pre-School in Object Identification

using Open CV

Object Information

Figure 5. DFD Level 0

In the figure above, there is only one user. The user needs to scan

the object to identify and analyze by the application which is the eTudlo: A

Mobile Leaning Tool for Pre-school in Object identification using Open CV and

immediately provide the information of the objects.

3.4.2.2 DFD Level 1

Figure 6. Illustration of the DFD Level 1

In the figure above, shows what the general movement is and the

way in which the application behaves when it is used. The first process after

the application has been opened; the object such as chair, board, table and etc.

Has been identify automatically.

In the second process, the information of the objects shows

automatically.
3.4.3 Implementation

During the implementation, the app function were develop. The syntaxes

and the queries needed to our app were created and build through the aid of

different software indicated in our Software Development Tools in Chapter 3.5

in order to make the app runnable.

3.4.4 Testing

After the implementation phase, proponents will now recognize the result

of application. This is to be done to recognize the original result and the

predictable result. Please see section 3.9 for the Verification, Validation and

Testing Plans.

3.4.5 Maintenance

The last phase is the maintenance. This phase can be work on after the

completion of the project. The proponents decided to monitor the system every

month. These are the types of maintenance that the proponents would possibly

go through.

 Corrective Maintenance: Correct errors which were not

discovered during the product development phase.

 Perfective Maintenance: Improve implementation of the

system. It enhances the functionalities of the system.

 Adaptive Maintenance: Port software to a new

environment.
3.5 Software Development Tools

3.5.1 Front-end (Tool)

 Android Studio – is the official IDE (Integrated Development

Environment) or tool for developing application exclusively for

android platforms.

3.5.2 Back-end (Tool)

 Python – The proponents would use this framework to have a

clean and simple routing. Python was a popular programming

language having simple, expressive and elegant syntax that

helps in creating an application.

 WHAT OPEN Computer Vision Library you are going to use.

3.6 Schedule and Time

August September October November

Activities 1 2 3 1 2 3 4 1 2 3 4 1 2 3 4

Planning

Designing

Coding

Testing

Maintenance
 The figure shows the timeline for the process and development of the

project.

3.7 Project Teams and Their Responsibilities

Table 5. Project Teams and Responsibilities

Team

Member Responsibilities Description Module

Name

Geralyn Team Provides a guidance, Admin

P. Leader/Researcher/ instructions, direction Module/

Masuhay Developer and leadership to a Mobile

group of other Learning in

individuals (the team) Object Identi-

for the purpose of fication

achieving a key result

or group of aligned

results. She will also be

assigned to program

other functionality of

the system.
 Diane Researcher/ Behave honestly and Analysts/

Rose G. Developer ethically in the course Mobile

Calzada of their research. Utilize Learning in

programming Object Identi-

principles, tools, and fication

technique to write

application codes.

Allvel P. Researcher/ Evaluate the potential Analysts/

Gorne Developer risks posed by the Mobile

research process. Learning in

Develop technical and Object Identi-

functional specification fication

for projects. Ensure

project deliverable

meet business

requirements.

3.8 Budget Cost Management Plan

3.8.1 Development Cost (Work Force)

Table 6. Development Cost (Work Force)

2 Months of

Function Salary Development

Team leader/Project 45,000 90,000

Manager
 System/Analysis 18,000 36,000

Developer 17,000 34,000

Developer 17,000 34,000

Total P97,000.00 P194,000.00

3.8.2 Hardware Cost

Table 7. Hardware Cost

Description Quantity Amount

Android Phone 1pc 2,999.00

Laptop 1pcs 18,739.99

Total P21,738

3.8.3 Facilities

Table 8. Facilities of the Development

Description Total Budget(2 Months

Development)

Work Cost 190,000.00

Electric Expense 4,000.00

Internet Expense 2,600.00

Printing Expense 2,000.00

Total P198,600.00

Table 8 shows the facility cost that the proponent has been used during the

development.
3.9 Verification, Validation and Testing Plan

This part contains the plan of the activities to be conducted with

developers and the direct beneficiary of the system.

References

Paul Henry “E‐learning technology, content and services”

Publication date: 1 June 2001. [Online]

Available: https://doi.org/10.1108/EUM0000000005485

Tosti H. C. Chiang, Stephen J. H. Yang and Gwo-Jen Hwang

Journal of Educational Technology & Society

Vol. 17, No. 4, Review Articles in Educational Technology (October

2014), pp. 352-365

Published by: International Forum of Educational Technology & Society

https://www.jstor.org/stable/jeductechsoci.17.4.352

Yueh-Min Huang, Yu-Lin Jeng and Tien-Chi Huang

Journal of Educational Technology & Society

Vol. 12, No. 2 (April 2009), pp. 163-175

Published by: International Forum of Educational Technology & Society

https://www.jstor.org/stable/jeductechsoci.12.2.163
 Language Learning & Technology s October 2013, Volume 17, Number

3 pp. 52–73

George, Ola.Wilmington University (Delaware), ProQuest

Dissertations Publishing, 2019. 13862048.

1. 1.

Talukder, Z., Basak, A., Shoyaib, M.: Human skin detection. J. Comput. Sci.

Technol. Graph. Vis. (2013)Google Scholar

1. K. Lai, L. Bo, X. Ren, D. Fox, "A large-scale hierarchical multi-view RGB-D

object dataset," IEEE International Conference on Robotics and Automation

(ICRA), pp. 1817-1824, 2011.

Show Context View Article Full Text: PDF (2020KB) Google Scholar

https://search.proquest.com

https://search.proquest.com/openview/a348827d4fc6aaf650f26a36d2e

92e28/1?pq-origsite=gscholar&cbl=18750&diss=y

J. Prankl, A. Aldoma, A. Svejda, M. Vincze, "RGB-D object modelling for

object recognition and tracking", Proc. IEEE/RSJ Int. Conf. Intell. Robots Syst.

(IROS), pp. 96-103, 2015.

Show Context View Article Full Text: PDF (4142KB) Google Scholar
Tan, Qing; Chang, William; Kinshuk

Electronic Journal of e-Learning, v13 n2 p138-148 2015

MobiSys '17 Proceedings of the 15th Annual International Conference on

Mobile Systems, Applications, and Services

Pages 82-95

Niagara Falls, New York, USA — June 19 - 23, 2017

ACM New York, NY, USA ©2017

table of contents ISBN: 978-1-4503-4928-4doi>10.1145/3081333.3081360

ASSETS '13 Proceedings of the 15th International ACM SIGACCESS

Conference on Computers and Accessibility

Article No. 20

Bellevue, Washington — October 21 - 23, 2013

ACM New York, NY, USA ©2013

table of contents ISBN: 978-1-4503-2405-

2 doi>10.1145/2513383.2513443

Sungkur, R., Panchoo, A. and Bhoyroo, N. (2016), "Augmented reality,

the future of contextual mobile learning", Interactive Technology and

Smart Education, Vol. 13 No. 2, pp. 123-

146. https://doi.org/10.1108/ITSE-07-2015-0017

[online] Available:

https://github.com/BVLC/caffe/tree/master/models/bvic_reference_caffen

et/.

[online] Available:

https://github.com/BVLC/caffe/tree/master/models/bvic_reference_caffen

et/.

Joseph Redmon, Santosh Divvala, Ross Girshick, Ali Farhadi; The

IEEE Conference on Computer Vision and Pattern Recognition (CVPR),

2016, pp. 779-788

https://doi.org/10.1111/bjet.12147

Mobileye Pedestrian Collision Warning System

, http://www.mobileye.com/en-uk/mobileye-features/pedestrian-collision-

warning/. (accessed 1 January 2016).

Google Scholar
Research Scholar, Department of Electronics & Communication,

University Institute of Technology, Rajiv Gandhi Technical University,

Bhopal – 462033 M.P. (India)

International Journal of Educational Technology in Higher

Educationvolume 15, Article number: 10 (2018)