Abstract | ||
---|---|---|
Counterfeit currency is a very serious issue from an economic point of view for any country. In rural communities around the world, where digital currency has not yet become the norm and the usage of banknotes is still dominant, the detection of counterfeit banknotes is very important. Authentication of currency is a classification problem and therefore a Fuzzy Inference System (FIS) is designed with its parameters optimized using a Genetic Algorithm. This FIS takes in features extracted from the images of banknotes and classifies the corresponding banknotes as genuine or fake. A cascaded approach is selected to design the FIS in order to reduce the number of parameters. In this study, an algorithm based on the Fuzzy C-means clustering algorithm is designed to define the best combination of inputs for the cascaded FIS. Also, another cascaded FIS is developed with an attempt at decreasing the number of parameters by introducing symmetry and other constraints. The results are compared with the Machine Learning (ML) approaches, and efforts are also made to get a highly reliable precision similar to that of ML methodologies. |
Year | DOI | Venue |
---|---|---|
2020 | 10.1007/978-3-030-81561-5_23 | FUZZY INFORMATION PROCESSING 2020 |
DocType | Volume | ISSN |
Conference | 1337 | 2194-5357 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Anirudh Chhabra | 1 | 0 | 1.01 |
Donghoon Kim | 2 | 0 | 1.69 |
Kelly Cohen | 3 | 0 | 1.01 |