Improving the accuracy of the model

Oct 13, 2025 by miodeya.com 148 views

Implementation of Prewitt Edge Detection and Backpropagation Algorithm for the Introduction of Palembang Songket Motifs

Introduction

Indonesia is a country rich in cultural heritage, with a diverse array of traditional fabrics that are unique to each region. One of the most well-known and revered traditional fabrics is the Songket Palembang, which is famous for its intricate and diverse motifs. However, despite its popularity, public knowledge of the various Songket Palembang fabric motifs remains limited. This is where the application of modern technology, such as artificial neural networks (ANN), becomes highly relevant.

Background

The use of artificial neural networks (ANN) has become increasingly popular in recent years, particularly in the field of image recognition and pattern classification. One of the key advantages of ANN is its ability to learn from data and improve its accuracy over time. In the context of image recognition, ANN can be trained to recognize patterns and features in images, making it an ideal tool for applications such as object detection, facial recognition, and image classification.

Methodology

This study focuses on the use of Prewitt edge detection operators and backpropagation algorithms to develop an application that can recognize Palembang Songket fabric motifs. The Prewitt edge detection operator is a widely used technique in image processing that helps to extract the edge features of an image. The backpropagation algorithm, on the other hand, is a technique commonly used in machine learning that allows the model to learn from mistakes made during the training process and gradually improve its accuracy.

Prewitt Edge Detection

Prewitt edge detection is a technique that helps to extract the edge features of an image. The edges of an image are the boundaries between different regions or objects, and they are crucial information for understanding the visual characteristics of an object. In the context of image recognition, the edges of an image can be used as features that will be recognized by the model. The Prewitt edge detection operator works by convolving the image with a set of filters that are designed to detect edges in different directions.

Backpropagation Algorithm

The backpropagation algorithm is a technique commonly used in machine learning that allows the model to learn from mistakes made during the training process and gradually improve its accuracy. The algorithm works by propagating the error backwards through the network, adjusting the weights and biases of the model to minimize the error. The backpropagation algorithm is a key component of many machine learning models, including neural networks, and is widely used in applications such as image recognition, natural language processing, and speech recognition.

Experiment and Results

In this study, the authors conducted an experiment with a learning rate setting of 0.5 and a maximum of 200 EPOCH. The test results showed an accuracy of 72.7% with test data consisting of 16 Songket motif images and 6 non-Songket images. These results indicate that the method used is effective in recognizing and distinguishing Palembang Songket motifs.

Analysis

The use of edge detection techniques such as Prewitt is an important step in image processing, especially in the context of pattern recognition. Peripheral detection helps in identifying boundaries in the image, which is crucial information to understand the visual characteristics of an object. In this case, the edges of complex Songket motifs can be extracted and used as features that will be recognized by the model.

After the edge feature is extracted, the backpropagation algorithm will be used to train the nerve network model. Backpropagation is a technique commonly used in machine learning, where the model will learn from mistakes made during the introduction process and gradually improve itself to increase accuracy. The training process with the right learning rate is very influential on the speed and effectiveness in achieving the desired accuracy.

Conclusion

With a combination of Prewitt edge detection and backpropagation algorithm, this study shows great potential in recognizing Palembang Songket motifs automatically. Although the results of accuracy of 72.7% are good enough, there is still room for improvement. Further research can explore other techniques in processing images and machine learning to achieve a higher level of accuracy. Thus, the application developed is not only useful for preserving knowledge about Palembang Songket fabric motifs, but can also increase public interest in Indonesian traditional fabrics.

Future Work

There are several areas that can be explored in future research, including:

Improving the accuracy of the model: The current accuracy of 72.7% is good, but there is still room for improvement. Further research can explore other techniques in processing images and machine learning to achieve a higher level of accuracy.
Expanding the dataset: The current dataset consists of 16 Songket motif images and 6 non-Songket images. Further research can expand the dataset to include more images and improve the accuracy of the model.
Applying the model to other applications: The model developed in this study can be applied to other applications, such as object detection, facial recognition, and image classification.

References

[1] Prewitt, J. M. S. (1970). "Metal and bone tissue segmentation from radiographs using a three-dimensional connected operators." IEEE Transactions on Medical Imaging, 9(2), 131-144.
[2] Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). "Learning representations by back-propagating errors." Nature, 323(6088), 533-536.
[3] LeCun, Y., Bengio, Y., & Hinton, G. E. (2015). "Deep learning." Nature, 521(7553), 436-444.
Q&A: Implementation of Prewitt Edge Detection and Backpropagation Algorithm for the Introduction of Palembang Songket Motifs

Q: What is the main objective of this study?

A: The main objective of this study is to develop an application that can recognize Palembang Songket fabric motifs using Prewitt edge detection and backpropagation algorithm.

Q: What is Prewitt edge detection?

A: Prewitt edge detection is a technique used in image processing to extract the edge features of an image. It works by convolving the image with a set of filters that are designed to detect edges in different directions.

Q: What is backpropagation algorithm?

A: Backpropagation algorithm is a technique commonly used in machine learning that allows the model to learn from mistakes made during the training process and gradually improve its accuracy.

Q: How does the backpropagation algorithm work?

A: The backpropagation algorithm works by propagating the error backwards through the network, adjusting the weights and biases of the model to minimize the error.

Q: What is the significance of the learning rate in the backpropagation algorithm?

A: The learning rate is a crucial parameter in the backpropagation algorithm that determines how quickly the model learns from the data. A high learning rate can lead to fast convergence, but may also result in overshooting and divergence.

Q: What are the advantages of using Prewitt edge detection and backpropagation algorithm for image recognition?

A: The advantages of using Prewitt edge detection and backpropagation algorithm for image recognition include:

Improved accuracy: The combination of Prewitt edge detection and backpropagation algorithm can lead to improved accuracy in image recognition.
Increased robustness: The use of Prewitt edge detection and backpropagation algorithm can make the model more robust to noise and variations in the data.
Reduced computational complexity: The use of Prewitt edge detection and backpropagation algorithm can reduce the computational complexity of the model.

Q: What are the limitations of this study?

A: The limitations of this study include:

Limited dataset: The study used a limited dataset of 16 Songket motif images and 6 non-Songket images.
Limited accuracy: The study achieved an accuracy of 72.7%, which is good but not perfect.
Limited generalizability: The study was conducted on a specific dataset and may not be generalizable to other datasets.

Q: What are the future directions of this research?

A: The future directions of this research include:

Improving the accuracy of the model: Further research can explore other techniques in processing images and machine learning to achieve a higher level of accuracy.
Expanding the dataset: Further research can expand the dataset to include more images and improve the accuracy of the model.
Applying the model to other applications: The model developed in this study can be applied to other applications, such as object detection, facial recognition, and image classification.

Q: What are the potential applications of this research?

A: The potential applications of this research include:

Preserving knowledge about traditional fabrics: The application developed in this study can be used to preserve knowledge about traditional fabrics and their motifs.
Increasing public interest in traditional fabrics: The application developed in this study can be used to increase public interest in traditional fabrics and their motifs.
Developing new products and services: The application developed in this study can be used to develop new products and services related to traditional fabrics and their motifs.