www.rsisinternational.org
Page 165
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
ESPYREAL: A Mobile Based Multi-Currency Identifier for Visually
Impaired Individuals Using Convolutional Neural Network
*De leon, Kurt Russel N., Datiles, Jhodel D., Cruz, Kyle Condrei N., Cruz Kathlyn M., Ronald
Fernandez
*College of Computing Studies, Universidad De Manila, Philippines
ABSTRACT
Visually impaired person’s faces many issues in handling money because they cannot differentiate different
denominations especially in countries like the Philippines where both local and foreign banknotes are circulating
together. This challenge may lead to various issues like the misidentification of paper money or coins, depending
on others for financial matters, and being open to possible exploitation of finances. Therefore in this study the
proponents implemented an image processing techniques that will assist visually impaired people in detecting
and identifying money using convolutional neural network algorithm. The main objective of this application is
to help the visually impaired individuals identify two different denominations that is commonly used in the
contemporary time of the Philippines such as United States Dollar (USD) and Philippine Peso (PHP), making
them feel secured and confident when they are conducting financial transaction alone. The application will be
implemented as Android-based money detection app. The researchers utilize the AGILE and Tensoflow platform
to build accurate and fast model. They collected diverse amount of Philippine Peso and United States Dollar
banknotes and coins images, captured in different angles and lighting condition to achieve reliable model for
Multi-Money Recognition Application. Furthermore, The proponents created a likert scale questionnaire that
will use for survey and interview with visually impaired stakeholders in Pasay, Manila. Based on ISO/IEC 25010
evaluation, Espyreal achieved excellent ratings across functionality (4.74), usability (4.72), performance
efficiency (4.64), reliability (4.60), and portability (4.58), with an overall weighted mean of 4.60. A 98 percent
accuracy across all bills is achieve through diverse collection of bills datasets and aggressive training with
Tensorflow platform. The results demonstrate that the system is functional, dependable, efficient, and user-
friendly. The proponents suggest to upload the Espyreal to the Google Play Store for Easy access and download
for the intended users.
Keywords: Tensoflow, Multi-Money Recognition Application, CNN.
INTRODUCTION
Nowadays, technology is growing fast. By the use of technology, people are able to solve their problems within
the small passage of time. Some applications are used in real life which includes currency monitoring systems,
currency counting machines, currency exchange machines and currency recognition systems to help for blind or
visually impaired people. Imaginative and perceptive no longer only facilitates us to carry out each day activities
however also impact the behavior of the individual. Normal people can easily recognize the currency, but it
becomes very difficult for the visually impaired to accurately recognize the currency. The blind or visually
impaired people need to recognize and also adapt to differences among the note currencies
Tom [1] writes in his article that blind or visually impaired people must be able to distinguish their currencies
swiftly and safely while paying and receiving change. In the U.S., all bills are the same size. Blind people can't
tell denominations apart. Banknote sizes are different in several countries, such as India, Australia, Malaysia and
Philippines, making money identification easier. Blind people may measure and identify money with a money
identity card. When a currency is lined up with the card, tactile markings determine which bill to use. However,
the blind individual needs help to determine money's value. Another way is blind people can fold
www.rsisinternational.org
Page 166
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
money for specific denominations. However, they may have trouble sorting through many denominations while
carrying multiple currencies. Tom said that identifying money varies by necessity and currency, and each choice
has pros and cons. If people could distinguish banknotes by touch, life would be easier, improving and securing
cash identification.
Therefore This study seeks to develop Espyreal: A Multi-Currency Identifier that will give a reliable, user-
friendly, and accessible tool to the visually impaired sector in the Philippines. Through technology, including
image identification and voice feedback, the system will be able to let the users properly identify banknotes by
themselves without anyone's intervention. The significance of this study lies in its contribution to realizing
financial inclusiveness, independence, and security of the visually impaired population, in compliance with the
higher objective of accessibility and equal opportunity for all. This research aims to give a dependable and
accessible solution that equips the visually impaired to handle money confidently and securely in day-to-day
transactions.
Related Studies
According to [2], In India, nearly 12 million visually impaired people had difficulty in identifying the currency
notes. Due to this problem the proponents in this study propose a solution called Coinnet. The proposed system
is developed using TensorFlow and improved by selection of optimal hyperparameter value, and compared with
existing well known CNN architectures using transfer learning. Based on the results it has been observed that
the proposed model performs well over six widely used existing architectures in terms of training and testing
accuracy.
A Study by Pathak, A. & Aurelia, S. [3] Stated that Around 37 million people Visually impaired individuals in
India face challenges in checking the currency due to the lack of Braille systems and the similar sizes of each
currency. To address this issue, a mobile-based Indian currency detection model is proposed, which uses image
processing for feature extraction and a basic CNN for currency identification. This model is being developed
into a mobile application, allowing visually impaired individuals to quickly check for potential frauds. The
results are promising, around 80% of the visually impaired are relying on this application to identify banknotes.
A study by Kumar, K., et al. [4] a special mobile application designed for support of visually impaired people
during identification of currency throughout transactions. By employing the Random Forest Classifier in
conjunction with OpenCV, it correctly identifies many different currency notes. It does so through machine
learning. The Text-to-Speech function tells people the identified amount, while also helping independence and
swift payment.
According to the study of Vivien Agustin, et al [5] A visually impaired individuals face challenges in recognizing
banknotes due to the similarity of paper texture and size between different categories. To address this, researchers
have developed an Android-based money detection app called Eyessential, which uses image processing. The
app is functional, reliable, helpful, efficient, maintainable, and portable. The developers suggest submitting the
Eyessential Android app to Google Play for easy installation and compatibility with Google Assistant for blind
users. A flashlight option is also proposed for user-friendliness. To accurately detect denominations, substantial
data sets are suggested.
In the study of J. A. P. De Jesus, et al [6] Visually impaired people must have tools, assistant devices, or
organizations that guide them throughout their lives but at a certain point, other visually impaired are lack of
material especially in learning. Researchers in this study believed that having an environment that supports
visually disabled people will serve as their great foundation throughout their lives.
Conceptual Framework
This part focuses on the structure framework of the application for better understanding and analyzing how the
application works, and how it processes gathered information.
www.rsisinternational.org
Page 167
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
Figure 1: input-process-output model (IPO)
Figure 1 illustrates how Espyreal application workflow using input-process-output model, in the input stage,
only needs a scanned currency, using the phone camera of the end-user to scan a banknote and coins.
The process stage consists of recognizing scanned currency and identifying banknotes & coins denomination
using Convolutional Neural Networks, and feedback using Text-To-Speech. To identify the scanned currency,
it needs to be compared by its key features such as patterns, color, size, etc. With the help of Convolutional
Neural Networks to recognize and identify the scanned currency denomination and its type of banknote whether
it is a Philippine Pesos and United State Dollars Then the system will start processing the gathered data and turn
it into audio feedback using text-to-speech, the system will announce the scanned currencies type and
denomination for more accurate currency identification. Additionally, the system will ask the end-user if they
want to add the current banknote to the next scanned currency for easy and smooth transaction.
And in the output stage the application will announce the scanned currency denomination, type, and the current
counted banknotes. Thereafter, the system will show a text display of all the gathered data of the application
ensuring a secure user experience. This application encourages and empowers individuals with visual
impairment to do transactions securely and confidently.
Application Development
For the application development, the proponents in this this adapt the AGILE Life Cycle Model due to the
reasons that the the agile model was primarily designed to help a project adapt quickly to change requests which
is suited for the application development and prototyping. The primary use of the Agile model in this study is
to facilitate quick project completion. To accomplish this task, it is important that agility is required. Agility is
achieved by fitting the process to the project and removing activities that may not be essential for a specific
project. As a result, the propronents can ensure that their target users are satisfied with the quality of their work.
Figure 2: Espyreal Prototype
Figure 2 shows how will the system work after the development, First the when the user first launch the
application it will directly go to the loading screen or splash-screen where all Philippine peso model and United
www.rsisinternational.org
Page 168
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
States Dollar Model will load during that time where in approximately 2-3 seconds. After that, the application
will immediately proceed to the identifying section where in the users have accessible button and medium view
camera, if the users click or push the capture button it will capture the image and passed to the model, the model
will identify what is the denomination of the bills and give the result to the screen with audio announcement.
The big buttons and easy navigation of the application was intentional because the main target of the application
is visually impaired, making the application user-friendly and reliable.
Requirements Gathering
The agile first phase is requirements. In this phase, the proponents must identify and gather the key requirements
needed to develop the project and understand what the customer or end-user needs by interacting with them
through surveys or interviews to make sure that everyone is on the same page.
This phase also holds the planning of how much time and effort the developers should give to assess if the project
is technically possible and whether it is worth the investment from both technical and economic standpoint.
Furthermore, The proponents must learn deep learning techniques such as Convolutional Neural Network (CNN)
for recognition and identification of the currencies and a subfield of Natural Language Processing (NLP) which
is for the text-to-speech feature for audio feedback which is required to build the Espyreal. And conduct a survey
and interview to identify the needs of the visually impaired when it comes to handling and managing money.
Additionally The proponents must gather and study different platforms and tools that will use in the development
of the application. In this phase the proponent decided what features of the application can help the visually
impaired and what features and functionality they will need to make the Money Identifier application more
accessible and usable to them.
Design
The agile second phase is design, in this step, the proponents will create a plan or design for the application or
system based on the gathered requirements. The proponents will use a user-flow-diagram to discuss how the
application works, it is shown in the Conceptual Framework figure number 1 the IPO or the input process output.
Figure 2 illustrates the appearance of the money identifier application. Once the app is open, it requires no input
from the user and directly navigate to the identifying screen, which will be easily accessible to the visually
impaired. The planned design of the user interface is done in this phase, this shows how the application will
look and how the end-users will interact with it, ensuring it is user-friendly and easy to navigate.
As for the user interface, the proponents designed a simple but interactive UI/UX that will meet the needs and
satisfaction of the end-users, and to do this, the proponents will use Figma for prototyping due to it being the
best option for designing and simulation of the application,
For the developing an interactive mobile-based application the proponents decided to use react-native to achieve
the modern and clean looking UI/UX. Espyreal is designed to be straightforward ensuring it is an easy to use
tool for the visually impaired.
Development
The agile third phase is the development, in this step, the developers will start developing the core components
of the application based on the gathered requirements and planned design. In this phase of the development
model happens the actual programming and building of the application to make it functional and accessible when
it moves on to the next phase.
The proponents start by establishing the needed datasets and build the foundation and core frameworks of
Espyreal such as the system algorithms, specifically the Convolutional Neural Network (CNN) that will classify
and identify the currencies and polish all errors that will occur in the application. This phase also holds the
creation of the user interface that will allow the end-users to use the functionalities of the application with ease
that will meet the needs and satisfaction of the visually impaired individuals.
www.rsisinternational.org
Page 169
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
The proponents use old phone camera and new phone camera to capture PHP and USD bills and coins, this
approach ensure that the result of the model is accurate on all devices whether the user phone is old and new.
After gathering the datasets is normalize in google-collab and apply medium to light augmentations to achieve
accurate results in real-world usage. The proponent’s will maximize the Teachable machine capabilities to train
the datasets because it allows the proponents to build and test the model in short period of time. Moreover the
proponent’s will use React Native to build the pipeline and image processing functions to ensure that the image
that will capture of the application is correctly passed down into the models. In addition the React Native will
be used to create the design, functionalities and user interface. These include the Loading screen, identification
section, Audio Announcement Integration and offline capabilities. The Espyreal Application will be developed
on Lenovo Ideapad Slime one 8GB ram Laptop with AMD Ryzen 3 5300U with Radeon Graphics.
Testing
The agile fourth phase is the testing, it is the most fundamental part of the agile model for the reason that iteration
development happens in this phase, ensuring that the quality of the product is at its best ensuring to deliver a
reliable product for the target audience of the study.
In this stage, the proponents conduct an iterative system testing to evaluate and review the system performance
and ask for a second opinion with the application target audience for any refinements.
After the application is done, the proponents will first test the model in the teachable machine testing area, where
the proponents feed the new images of banknotes and coins into the model to test its accuracy and once it is
accurate in the teachable machine, the model is exported and eventually integrated into the application where in
the proponents will test the model on the banknote and coins from different conditions, if there is any color
shifting, and if there are any changes to the quality of the banknote, and as for the coins the proponents will
check the size and if there is deformation,
Moreover, the proponents prepare a likert scale interview questionnaire for the intended users in Philippine
national School for blind in Pasay to evaluate and test the application, the questionnaire is also based on the ISO
9126 criterion to make sure both models and application satisfy the user needs and demands.
This study uses purposive sampling to select visually impaired participants based upon their own availability
and willingness for participation. One hundred Fifty (150) visually impaired individuals in the Philippine
National School for blind were selected to test the functionalities and accuracy of the application.
Deployment
In the fifth phase is the deployment part, in agile this phase where finished software is made available to end-
users in a production environment. It is a crucial phase due to the reason that it allows the developers to get
feedback from users. This phase of agile methodology ensures that the software or application is stable and ready
for users.
After the testing and evaluating the system performance and checking any vulnerable part of this system
application, the proponents will deploy it first using a link provided by the react native platform, the intended
users can download application through qr code or link, if is easy to install, once the user download the apk in
the link, he/she can install it immediately to identify both USD and PHP. The proponents suggest to submit the
application in the google play store so that visually impaired can download it, anywhere and anytime.
Additionally, the users are required to have internet connection they want to download it, but once it installed,
the functions and identification that are available in the app is offline.
Review
In the last stage, review is the last phase of the agile methodology. In this phase, the developers will review the
process helping ensure that the team and stakeholders are aligned on the product vision and objectives.
After the deployment the proponents will review system performance based on the target audience feedback,
www.rsisinternational.org
Page 170
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
and make any changes to the application to ensure stability of the system functionality and adaptability, meeting
end-users satisfaction.
Espyreal Algorithm
Figure 3 : Dataflow Diagram of Espyreal
The Data Flow Diagram (DFD) of the Espyreal Application as shows in the figure 3, shows the application flow
and process under the hood, These process involved several steps such as scanning and identifying the money
of the visually impaired user once it captures image. The algorithm will start running once the visually impaired
user scan the money, the application will detect the images and feed the images into the model. After that the
Tensorflow.js model will process the image and try to identify the image, once the model correctly identify the
image, it will give the result back to the application. Then the application will start process that result into the
User interface of the application with loud and clear voice announcement. It displays how the applications
identify the money of visually impaired users under the hood.
Figure 4: Generated Sample Money Identification (a)-(e) True Positive Results (f) True negative.
RESULTS AND DISCUSSION
Philippine and United States Dollar Recognition
The developers have created a simple yet modern GUI or Graphical User Interface. The use of the Espyreal
Application is simple and straight forward, once the application runs, it is ready to identify the bills and required
www.rsisinternational.org
Page 171
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
no input to the visually impaired users, making the application both accessible and easy to use. Multiple test
have been implemented in the testing phase of the application, the bills is captured in every direction and angles
visually impaired might identify. As shown in figure 6. In addition the figure clearly shows the equivalent
percentage in accuracy form. As shown, the results is promising, from (a)-(e) the results shows 99% percent of
accuracy. Although the application is challenge if the bill is far way from the camera. Furthermore in order for
the model to have accurate results, at least half of the bill should be seen in the camera. In addition the application
allow the users to scan the bills in natural, dim, and dark lighting due to the capability of the application to enable
the flashlight or torch functions on the phone. The proponents were able to achieve (90-99%) of accuracy on
both PHP and USD bill datasets especially on the 20pesos, 50pesos, 100pesos, 200pesos, 500pesos, 20 dollars,
5 dollars, 1 dollars, 50dollars, 100 dollars. Moreover the application is limited only to the USD and PHP bills
and coins due to the datasets and financial limitations.
Evaluation based on ISO 9126-1 criterion
A survey of 150 visually impaired individuals in Philippine national school for blind was conducted, sixty two
(62) of them are utterly blind, eighty eight (88) of them are partially blind. The ISO 9126-1 survey questions
were created to evaluate if the application actually help them to identify common currencies that circulating in
the Philippines and also if the application make their day-to-day transaction safe and easier. In addition by
creating this questions it also allows the proponents to gather feedbacks and improve the application catering
their needs and satisfaction. Goggle forms were use to analyze and quantified the data gathered answers from
the visually impaired participants, ensuring the data is clear and accurate.
Functionality
Functionality refers to the application's usefulness or efficiency. The app received 4.74, 4.74, 4.76, 4.72, and
4.58. Mean scores for criteria 1 to 5, respectively, as clearly shown in the table below
Table 1: User Evaluation Summary
1. Overall, the weighted mean is 4.60, which is equivalent to Excellent. These results indicate that the majority
of respondents found the application highly functional, efficient, reliable, and user-friendly.
www.rsisinternational.org
Page 172
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
Performance
The system gained a mean of 4.64 (Excellent). Participants highlighted that the application identified currencies
quickly enough (less than one second) for practical real-world use while consuming minimal. device resources
such as battery and memory.
Figure 5: Bar Graph of Performance Criteria
2. The figure 5 clearly shows the the summary of the answers of the users based on their experience while using
the application. This confirms that Espyreal is not only accurate but also efficient in execution, even on mid-
range smartphones.
Compatibility
3. Compatibility refers to the quality of the Eyessential money identifier application's ability to work effectively
alongside other accessibility applications and integrate seamlessly with device accessibility settings. The first
criteria under Compatibility received a weighted mean of 4.58, which corresponds to Excellent, as shown in
Table 1. The weighted mean for the second criteria was 4.54, also achieving an Excellent rating.
According to the respondents, the application works well alongside other accessibility applications without
conflicts or interference. Users reported that the app integrates properly with their device's existing accessibility
settings and features. The application demonstrates strong compatibility with various assistive technologies that
visually impaired users commonly rely on. Respondents appreciated that they could continue using their other
essential accessibility tools while benefiting from the currency identification features..
Figure 6: Bar Graph of Compatibility Criteria
4. Users expressed confidence in the application's ability to function as part of their broader accessibility toolkit.
With an average mean of 4.56, as shown in figure 6. Most respondents assessed the Espyreal application as
www.rsisinternational.org
Page 173
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
excellent in terms of Compatibility, indicating that the app successfully meets the interoperability needs of
visually impaired users in their daily digital environment
Usability
Usability refers to the ease of use and user experience quality of the Eyessential money identifier application for
visually impaired users. The criteria under Usability achieved weighted means of 4.75, 4.74, 4.72, and 4.70, all
corresponding to Excellent ratings, as shown in Table 1.
Figure 7: Bar Graph of Usability Criteria
5. Respondents found the application's interface simple, appealing, and straightforward to understand. Users can
operate the app with minimal effort and reported that navigation is intuitive for the target population. The app is
easy to learn and use effectively for first-time users, with clear audio feedback and logical gesture-based controls.
Reliability
Reliability refers to the ability of the Eyessential money identifier application to perform consistently and
maintain stable functionality over time. The criteria under Reliability achieved weighted means of 4.62 and 4.60,
both corresponding to Excellent ratings, as shown in Table 1.
Based on the respondents answers the application rarely crashes and performs consistently during use. Users
found the app to be dependable and always available when needed for currency identification tasks. The
application maintains stable performance across extended usage periods without significant errors or
interruptions.
Figure 8: Bar Graph of Reliability Criteria
www.rsisinternational.org
Page 174
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
6. As shown in figure 8, in terms of Reliability the application 4.60 (Excellent) reflecting that Espyreal was
always available when needed and was able to recover smoothly from interruptions such as temporary
permission issues.
Security
Espyreal obtained a mean of 4.52 (Excellent). Although slightly lower compared to other criteria, this still falls
under Excellent.
Figure 9: Pie Chart of Security Criteria
6. Respondents acknowledged that the application protected their privacy since no sensitive data was collected
and all detection processes were performed locally. This offline-first design reassured participants that their
usage was secure and confidential
Maintainability
Although it was rated at the lowest among the other categories, (4.48). Respondents still recognized that while
the system can be updated and adapted to new currencies, there remains room for improvement in providing
clearer diagnostic information and feedback
Figure 10: Bar Graph of Maintainability Criteria
7. Nevertheless, the modular design of Espyreal ensures that updates and improvements can be implemented
without disrupting existing features.
Portability
8. Espyreal was successfully tested across different Android versions (9 to 14) and phone models, showing strong
portability. With a mean of 4.58, showing that the system works effectively across different smartphone models
and operating system versions.
www.rsisinternational.org
Page 175
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
Figure 11: Pie Chart of Portability Criteria
8. Participants found the application easy to download, install, and set up, especially with audio-assisted
guidance during installation.
Data Availability
The datasets generated and analyzed during this study are not publicly available due to participant privacy and
confidentiality considerations involving visually impaired individuals from the Philippine National School for
the Blind.
The survey responses contain sensitive information about participants' accessibility needs and experiences.
However, anonymized summary statistics and evaluation results are presented in this manuscript.
Raw image datasets of Philippine Peso and United States Dollar banknotes used for model training are not shared
to prevent potential misuse for counterfeiting purposes. Aggregated data supporting the study's conclusions may
be available from the corresponding author upon reasonable request and with appropriate institutional approval.
Ethical Consideration
Informed Consent: Espyreal ensures that all users, especially those with visual impairments, understands the
reason why the study is conducted, how it was done, and see if there are any possible risks or benefits by using
the app.
Accessibility: Ensures that users can easily access the Espyreal regardless of the level of their visual impairment.
It is also accessible for users regardless of their age, technical proficiency, and educational background. All of
these ensure that the Espyreal system is accessible regardless of the status of the user.
Reliability and Accuracy: The app ensures the stability and consistency of the system, minimizing errors and
inaccurate results. A clear reminder should inform users of the possible identification errors and encourage
verification when necessary.
Fairness: Espyreal recognizes every banknote regardless of the condition, whether it is old, worn-out, or
crumpled. Ensuring that the recognition is accurate for all users and banknotes.
Transparency and Accountability: Espyreal informs users that it does not recognize counterfeit bills and should
not be taken upon for financial transactions. The users will have an easy way to report bugs or errors, considering
it for future improvements.
User Consent: The users do not have the option, thus not required to provide personal data to use the full
functionality of the app, Espyreal is designed to be accessible to all users with visual impairments without any
difficulties.
www.rsisinternational.org
Page 176
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN APPLIED SCIENCE (IJRIAS)
ISSN No. 2454-6194 | DOI: 10.51584/IJRIAS |Volume X Issue IX September 2025
CONCLUSION
Based on the findings, Espyreal successfully met its objectives of providing a reliable, fast, and accessible
solution for currency identification among visually impaired users. The integration of Convolutional Neural
Networks (CNNs) with audio feedback ensured accurate recognition and improved financial independence. The
high evaluation scores across all quality attributes demonstrate that Espyreal is not only functional and efficient
but also adaptable for future updates, making it a practical tool for real-world application. The study concludes
that Espyreal is a dependable assistive technology that can significantly improve the financial autonomy and
security of visually impaired individuals.
RECCOMENDATIONS
1. Expand Currency Support – Extend the model to include additional currencies commonly encountered by
Filipino users to enhance versatility.
2. Counterfeit Detection – Incorporate security features such as watermark or UV detection modules to safeguard
against counterfeit bills.
3. Broader Testing – Conduct larger-scale user testing in multiple regions to improve generalizability and
inclusivity of results.
4. Continuous Updates – Regularly update the system to accommodate new banknote designs released by central
banks.
5. Future Research – Explore the integration of haptic and auditory cues for a multimodal accessibility
experience, further improving usability for different levels of impairment.
ACKNOWLEDGMENT
The authors of this study would like to express their gratitude to the College of Computing Studies Department
in Universidad De Manila, for allowing them to conduct this study
REFERENCES
1. Tom, “How Blind People Identify Paper Money - Blind Coin Collector,” 2019.
2. Veeramsetty, V., Singal, G., & Badal, T. “Coinnet: platform independent application to recognize Indian
currency notes using deep learning techniques”, (2020
3. Pathak, A. & Aurelia, S “Mobile-Based Indian Currency Detection Model for the Visually Impaired”, (2020.
4. Kumar, K., et al. (2023) "Notepal,"
5. Vivien Agustin, et al (2022) entitled “Money Identifier: An Android-Based Application for Visually
Impaired People Using Object Detection.”
6. J. A. P. De Jesus, et al. (2021) entitled "PerSEEption: Mobile and Web Application Framework for Visually
Impaired Individuals,"
