SUMMARY
The use of biometrics in various issues of data security, transactions, privacy has been widely used. Regarding security, surely the level of accuracy in one's self-recognition must be accurate. The palmprint is one of biometrics which has various unique characteristics. These characteristics include geometrical features such as length, width, and area of the palmprint, characteristics of main lines such as heart lines, headlines, and lifelines, tangled or weak lines, delta dot characteristics, and minus characteristics. Main lines and tangled lines, which are often referred to as line features, have several advantages over other biometric features, such as they can be obtained from low-resolution images, are difficult to duplicate, and are stable because they have experienced slight changes over a long period of time. This research introduces the method of Fermat's Spirals in the extraction of the features of the palmprint which are represented in the time series curve. Testing uses 450 palm images sourced from 50 people, each person taken 9 palmprint images where 5 images as a test image and 4 images as a reference image. The test is carried out based on the threshold value determined from the simulation results based on matching test images with reference images on the ROC graph. Based on the test results it can be concluded that the system shows high performance at a threshold value of 0.036 to get an FNMR value of 2.80%, FMR of 0.88% with an accuracy of 96.31% where the number of reference images to the test image affects the level of accuracy.