ABSTRAKSI: Penggunaan informasi citra digital melalui jaringan telekomunikasi Internet meningkat secara signifikan. Proses pengiriman informasi citra digital melalui jaringan telekomunikasi membutuhkan bandwidth dan media penyimpanan yang cukup besar. Untuk mengatasi keterbatasan bandwidth dan media penyimpanan, maka perlu dirancang suatu sistem kompresi citra yang memiliki performansi yang optimal.
Pada tugas akhir ini, dilakukan penelitian mengenai implementasi kuantisasi Block Prediction Vector Quantization (BPVQ) pada domain spasial dan domain frekuensi berbasis transformasi DCT. Sedangkan, metode entropy encoding yang digunakan yaitu metode pengkodean Huffman. Citra digital yang digunakan adalah citra grayscale dengan format bitmap. Untuk mengetahui performansi kinerja sistem, maka dilakukan perbandingan dengan Kuantisasi Vektor.
Hasil yang didapat setelah pengujian, diperoleh rata-rata nilai rasio kompresi 46.20 %, rata-rata nilai PSNR 38.51 dB, waktu kompresi 27.1 detik dan waktu rekonstruksi 1.5 detik pada kuantisasi BPVQ pada domain spasial. Sedangkan, kuantisasi BPVQ pada domain frekuensi menghasilkan rata-rata nilai rasio kompresi sebesar 66.97%, rata-rata nilai PSNR 38.91 dB, waktu kompresi 29.5 detik dan waktu rekonstruksi 2.1 detik.Kata Kunci : Domain Spasial, Domain Frekuensi, Block Prediction Vector Quantization(BPVQ), Entropy Encoding Huffman, Transformasi DCTABSTRACT: The usage of digital image information trough the Internet telecommunication network increased significantly. The digital image information transmission process by telecommunication network needs large amount of bandwidth and storage. To resolve the limitation of bandwidth and storage problem, then it needs to be designed an image compression system that has the optimal performance.
In this final project, coducted research about the implementation of quantization Block Prediction Vector Quantization (BPVQ) in the spatial domain and frequency domain based on DCT transformation.While, the entropy encoding method that be used is Huffman encoding method. Digital image that be used is grayscale image with bitmap format. To find out the system's performance, then the comparison is used with Vector Quantization.
The result obtained of testing is the average value of compression ratio is 46.20%, the average PSNR value is 38.51 dB, compression time is 27.1 seconds and reconstruction time is 1.5 seconds on BPVQ quantization in spatial domain. While, BPVQ quantization in frequency domain obtain the average of compression ratio 66.97%, the average of PSNR is 38.91 dB, compression time 29.5 seconds and reconstruction time 2.1 seconds.Keyword: Spatial Domain, Frequency Domain, Block Prediction Vector Quantization(BPVQ), Huffman Encoding, Discrete Cosine Transform