Hasil Pencarian  ::  Simpan CSV :: Kembali

Abstrak :
ABSTRAK
Sistem pendeteksian wajah pada citra telah berkembang pesat sampai saat ini. Tujuan dari deteksi wajah adalah untuk mengindentifikasi dan menempatkan wajah manusia dengan pasisi. skala,oarientasi dan kondisi pencahayaan tertentu. Berbagai metode telah diajukan sampai saat ini. Salah satu pengembangan lebih lanjutnya adalah dengan menggunakan jaringan syaraf tiruan (neural network). Pada paper ini dibahas sistem deteksi wajah berdasarkan jaringan syaraf tiruan dengan metode training propagasi balik dengan momentum. Jaringan syaraf tiruan menguji setiap window dari citra, dan memmtukan apakah setiap window berisi wajah atau tidak. Setelah itu sistem menentukan window terbaik, yang akan disimpulkan sebagai wajah. Sistem inl dapat mendeteksi wajah frontal pada citra grayscale dengan latar belakang yang kompleks dan skala yang bervariasi. Agar dapat menguji citra masukan untuk ukuran wajah yang berbeda-beda, maka dilakukan metode piramida terhadap citra masukan.

Pada skripsi ini, ststem deteksi dengan jaringan syaraf tiruan diuji dengan perubahan pada parameter jumlah lapisan tersembunyi dan jumlah epoch yang dilakukan pada proses training. Sistem akan dianalisa kinerjanya berdasarkan lamanya waktu deteksi serta ketepatan hasil proses deteksi. Dari hasil pengujian didapatkan waktu deteksi sangat dipenganthi oleh ukuran citra, dan ketepatan proses deteksi sangat dipengaruhi oleh jumlah lapisan tersembunyi dan banyaknya epoch pada proses training, serta karakteristik dari citra masukan

---

Abstrak :
The Lanczos and conjugate gradient (CG) algorithms are fascinating numerical algorithms. This book presents the most comprehensive discussion to date of the use of these methods for computing eigenvalues and solving linear systems in both exact and floating point arithmetic. The author synthesizes the research done over the past 30 years, describing and explaining the "average" behavior of these methods and providing new insight into their properties in finite precision. Many examples are given that show significant results obtained by researchers in the field. The author emphasizes how both algorithms can be used efficiently in finite precision arithmetic, regardless of the growth of rounding errors that occurs. He details the mathematical properties of both algorithms and demonstrates how the CG algorithm is derived from the Lanczos algorithm. Loss of orthogonality involved with using the Lanczos algorithm, ways to improve the maximum attainable accuracy of CG computations, and what modifications need to be made when the CG method is used with a preconditioner are addressed.

Abstrak :
Many previous researches conveyed the superiority of Steepest Ascent (SA) method to find the optimal area in Response Surface Methodology (RSM) by shifting the experiment factor level. By using this method, Design of Experiment (DoE) is enabled to shift the factor level gradually in the right track, so that the global optimum can be reached. However, the response variable that is commonly optimized by using RSM cannot fulfill the classical statistics assumption of surface regression model. Taguchi’s orthogonal array, as alternative of RSM, gives loose statistics assumptions in performing the analysis. However, Taguchi’s orthogonal array has not yet been supported to shift the factor level to an optimum direction. Adopting the procedures of RSM in finding the optimal level combination using SA, integrating SA method in the Taguchi experiment is proposed in this paper. This procedure is applied into a simulated response surface. Then, the performance of this procedure is evaluated based on its direction to reach the optimum solution. The simulation data representing the real case is generated for two factors. Then, the proposed procedure is applied. The result of this simulation study shows that the integrated SA method in the Taguchi experiment successfully found the factor level combination that yields optimum response even though it is not as close as possible as the RSM results.