LED Control Using Hand Gestures
Jaspreet Kaur, Khushi Gupta, Shreya
ABSTRACT
This study presents a methodology for creating a system
that utilizes a model train set to control Light Emitting
Diodes (LEDs) based on real-time analysis of a physical
train layout. The process involves several key steps,
including image scanning to gather visual data, dataset
creation by compiling scanned images labeled with train
component information, model training to recognize
patterns and features within the images, real-time input
and prediction using the trained model on live video feed
frames, output translation into LED control instructions,
and LED activation based on model predictions. The
creation of the dataset involves image acquisition,
labeling, compilation, preprocessing, splitting, and
documentation to ensure robust model training and
evaluation. This approach facilitates the development of
an effective system for dynamically controlling LEDs in
synchronization with the actions of a model train layout.
INTRODUCTION
In our project, we embark on a pioneering endeavor
at the intersection of gesture recognition and LED
.
control, harnessing the power of machine learning to
enhance accuracy and responsiveness. With the fusion
of cutting-edge technologies, we aim to create an
intuitive system where users can control LEDs
through natural hand gestures, revolutionizing the
way we interact with our environment.
At the heart of our project lies the utilization of
machine learning algorithms to interpret and classify
hand gestures, enabling precise and real-time control
of LED illumination. By developing a robust model
Collection
trained on a diverse dataset of hand gestures, we seek
to achieve unparalleled accuracy in gesture
recognition, ensuring seamless and intuitive
interaction with the LED system.
Through this project, we not only aspire to push the
boundaries of human-computer interaction but also to
inspire innovation in the realm of smart lighting
solutions. By empowering users to effortlessly
manipulate their surroundings through simple
gestures, we envision a future where technology
seamlessly integrates into our daily lives, enhancing
both functionality and aesthetics. Join us on this
journey as we illuminate the path towards a more
intuitive and connected world.
RESEARCH POSTER PRESENTATION DESIGN © 2022
www.PosterPresentations.com
Manipal University Jaipur, Rajasthan, India
Dataset & Methodology
Image Scanning: Acquire images of real trains or train
components using a scanner or camera.
Dataset Creation: Compile scanned images into a dataset,
labeling them with relevant information such as train
component type (e.g., locomotive, wagon) and action
(e.g., moving forward, stopping).
Figure 1: Dataset Collection
In this figure, we are collecting a dataset comprising
images of fingers numbered from 1 to 10, captured using a
camera. We have used this dataset to train a machine
learning model, enabling accurate recognition of finger
gestures for diverse applications like sign language
translation or gesture-based interfaces.
Dataset Output
Model Predict
Fig2 : Proposed Methodology
Methodology Cont..
Model Training: Train a machine learning model using the
created dataset to recognize patterns and features within
the images.
Real-Time Input and Prediction: Apply the trained model
to analyze live video feed frames of a physical model train
layout, predicting identified components and actions.
Output to LED Control Code: Translate model predictions
into a format understood by LED control code, generating
specific signals or instructions based on identified
components and actions. LED Activation: Implement LED
control code to receive instructions from the model and
activate LEDs accordingly, reflecting the actions of the
model train layout.
Figure 3: Efficient Net Model
Controlling LEDs with Arduino involves connecting the
LEDs to specific pins on the Arduino board using wires to
regulate current flow. In the Arduino IDE, code is written
to define these pins as output and specify instructions for
LED control using hand gestures. After uploading the
code to the Arduino board, thorough testing ensures
proper functionality of the LED control system. Once
validated, the system is integrated with external inputs,
such as a model train set or sensors, allowing the LEDs to
respond dynamically to environmental stimuli. Finally, the
integrated system is deployed, ensuring all connections
are secure and the LEDs operate as intended in real-world
applications.
Result
In this system, the interaction between a machine learning
model, Python code, and an Arduino board enables the
control of LEDs based on detected hand gestures. Model
Prediction: The machine learning model analyzes input data,
likely hand landmarks from video frames, to predict hand
gestures .LED Control Function: The predicted output from
the model, such as the number of fingers detected, is sent to
the led(total) function in the Python code. Arduino Control:
Within the led(total) function, specific LED pins on the
Arduino board are manipulated using the write method from
the pyfirmata library. Each LED pin corresponds to a
particular finger or gesture, and its state (on/off) is
determined by the model's prediction. Physical Response:
The LEDs connected to the Arduino board respond
physically to the commands sent by the Python code. As a
result, they illuminate or turn off, providing a visual
representation of the detected hand gesture. Overall, this
system demonstrates a seamless integration of machine
learning, Python programming, and hardware control
through the Arduino board, enabling real-time interaction
and visual feedback based on detected hand gestures.
Conclusion
In conclusion, this project successfully demonstrates the
feasibility of controlling LEDs with hand gestures. The
developed system showcases the potential of hand
gesture recognition technology for creating a more
intuitive and interactive way to interact with electronics.
While the current prototype focuses on fundamental
functionalities, the future holds immense possibilities for
expansion. By incorporating more complex gesture
recognition, advanced LED control features, and
integration with other devices, this technology has the
potential to revolutionize human-computer interaction in
various domains, from smart homes and virtual reality to
assistive technology applications.
Overall, this project serves as a stepping stone towards a
future where hand gestures become a natural and
ubiquitous way to control the world around us.
Reference
1. Dayanandam, G.; Reddy, E.S.; Babu, D.B. Regression algorithms for efficient detection
and prediction of DDoS attacks. In Proceedings of the 2017 3rd International Conference on Applied and
Theoretical Computing and Communication Technology (iCATccT), Tumkur, India, 21–23 December 2017;
pp. 215–219.
2. Sharma, N.; Mahajan, A.; Mansotra, V. Machine Learning Techniques Used in Detection
of DOS Attacks:A Literature Review. Int. J. Adv. Res. Comput. Sci. Softw. Eng. 2016, 6, 100.
3. Somani, G.; Gaur, M.S.; Sanghi, D.; Conti, M.; Buyya, R. DDoS attacks in cloud
computing: Issues, taxonomy, and future directions. Comput. Commun. 2017, 107, 30–48.
4. Perera, P.; Tian, Y.-C.; Fidge, C.; Kelly, W. A Comparison of Supervised Machine
Learning Algorithms for Classification of Communications Network Traffic. In International
Conference on Neural Information Processing; Springer: Cham, Switzerland, 2017; pp. 445–454.