Hasil Pencarian  ::  Simpan CSV :: Kembali

"Salah satu kemampuan dari komunikasi multimedia adalah melakukan proses pengiriman gambar. Dalam proses tersebut akan sangat banyak membutuhkan ruang pengingat serta kanal transmisi. Pengembangan teknologi telekomunikasi dewasa ini lebih banyak ditujukan untuk lebih mengefisienkan penggunaan ruang pengingat dan saluran (kanal) transmisi yang dipakal. Semakin efisien penggunaan ruang pengingat maka akan semakin meningkat kecepatan pengolahan data dan selanjutnya akan meningkatkan kecepatan pengiriman gambar tersebut. Salah satu cara yang ditempuh untuk meningkatkan efisiensi penggunaan pengingat tersebut adalah dengan memampatkan gambar tersebut sebelum dilakukan proses selanjutnya. Dalam tulisan ini dijelaskan dan dibahas salah satu cara untuk mengefisienkan data image tersebut dengan menggunakan bantuan arsitektur jaringan saraf jenis MLP. Jaringan MLP ini dapat digunakan sebagai salah satu altematif untuk pemampatan data gambar dikarenakan kesederhanaan arsitektur dan imptementasinya. Salah satu tahap yang penting dari keseluruhan proses yang terjadi dalam jaringan MLP ini adalah tahap penentuan bobot-bobot hubungan node-node (simpul) jaringan MLP. Proses penentuan bobot-bobot yang biasa disebut sebagai proses learning dan suatu jaringan saraf, akan sangat berpengaruh terhadap sifat (chi) dan unjuk kerja keseluruhan jaringan MLP. Jika konfigurasi berubah maka secara keseluruhan sifat (ciri) maupun unjuk kerja jaringan tersebut akan berubah pula. Dua buah teknik learning yang banyak digunakan dalam proses learning (optimasi) adalah dengan memanfaatkan konsep backpropagation (BP) (metoda steepest descent) dan konsep Levenberg-Marquadt (LM) (salah satu pendekatan metoda Gauss-Newton). Tulisan ini menyajikan perbandingan unjuk kerja dua buah metoda tersebut dalam proses learning jaringan MLP serta gambar recall yang dihasilkannya. Hasil simulasi program menunjukkan bahwa metoda LM memberikan hasil recall gambar yang lebih bagus (high fidelity Irma recall kecil) serta mempunyai kecepatan konvergensi yang lebih tinggi jika dibandingkan metoda BP (SD).

---

One of the capabilities of the multimedia communication is the image communication. With this service, we will use so many memory spaces and transmission tines. Many of the development in telecommunication technologies are aimed to have the higher efficiency in using the memory spaces. The more efficient we use the memory spaces, the higher speed of the image data processing, and finally we can increase the speed of the image transmission.

One of the technique to have more efficiency in using the memory spaces is by compressing the image data before we stored and transmitted them. In this writing, we discussed one of the techniques for solving the image data compression, especially for still picture, by using the architecture of the neural networks. One of those types for this purpose is Multi Layer Perceptron (MLP). The MLP was choice in this purpose because of the simplicity of the architecture and the implementation.

One of the credal stages of the whole process in the MLP network is configuring the link weights of the nodes in the MLP network. Since the link weights are actually internal parameters associated with each node, changing the weights of the node will alter the behavior of the whole MLP network and then will change the MLP's performances.

There are two methods that often have been used in the learning process, the one was backpropagation (BP) method (steepest descent) and the other one was Levenberg-Marquadt (LM) concept (one of the approximations of the Gauss-Newton method). This writing is presenting the comparison between the performances of the LM method to the BP method in the learning process also the results of the recall pictures. The simulation shows that LM method gives the better image recall (higher fidelity) and has the higher convergence speed in the learning process."

"This book presents on the latest research findings, and innovative research methods and development techniques related to the emerging areas of broadband and wireless computing from both theoretical and practical perspectives.Information networking is evolving rapidly with various kinds of networks with different characteristics emerging and being integrated into heterogeneous networks. As a result, a number of interconnection problems can occur at different levels of the communicating entities and communication networks’ hardware and software design. These networks need to manage an increasing usage demand, provide support for a significant number of services, guarantee their QoS, and optimize the network resources.The success of all-IP networking and wireless technology has changed the way of life for people around the world, and the advances in electronic integration and wireless communications will pave the way for access to the wireless networks on the fly. This in turn means that all electronic devices will be able to exchange the information with each other in a ubiquitous way whenever necessary."