Tugas Akhir Informatika
Temu Kembali Citra Menggunakan Metode Modifikasi Multi Texton Histogram (MTH) dan Local Binary Pattern Rotation Invariant (LBPROT)
Image retrieval is a method used to search digital images in image database. One of the most important things in image retrieval is the feature extraction process. There are several methods that can be used in the image feature extraction process, one of which is Multi Texton Histogram (MTH). MTH is carried out with two main processes, namely color quantization and texture orientation. To improve the performance of the MTH method can be done by adding texton. In the testing process testing is carried out using rotated images. Therefore, the study also uses an invariant method for rotation, namely the Local Binary Pattern Rotation Invariant (LBPROT) method. The results of combining these two methods produce features of each image. The features produced by each image are compared using distance matrix. A query image that has the smallest distance matrix value is image that has the same class as the compared database image. Based on the results of testing on normal images, the use of Modified MTH and LBPROT can improve image retrieval performance by increasing the percentage of precision and recall by 3.63% and 0.62%. Whereas, in testing using rotated images, the merger of Modified MTH and LBPROT resulted in an increase in precision at angles rotation of 90 °, 180 °, and 270 ° by 2% -3%. However, The Modified MTH is better when compared to the merging of the Modified MTH and LBPROT on 45 °, 135 °, 225 °, and 315 ° angle rotations with precision values by 24% -27%.
2018179 | 006.68.Ash.t | Tersedia |
Tidak tersedia versi lain