1 対人プラスチック地雷可視化システム
廣瀬 明・増山 惣一・中尾一貴
2 干渉型開口合成レーダ
廣瀬 明・山城遼
3 適応的で高機能な超音波イメージング
廣瀬 明・西野智弘
4 光通信における適応的信号・情報処理
廣瀬 明・谷澤
5 複素ニューロCGHによる3次元動画の高速作製
廣瀬 明・谷澤 健・肥後智昭
6 遅延が学習にもたらす影響の解析
廣瀬 明・東 佑輔
7 画像中の文字抽出と処理
廣瀬 明・小山純平
8 神経細胞活動の定量的解析
廣瀬 明・渡辺 将也・森永真人
1 Antipersonnel Plastic Landmine Visualization System by Use of Complex-Valued Neural Networks
A.Hirose, S.Masuyama and K.Nakao
We focused on the detailed design of element antenna and the integration of functional modules. Last year, we proposed a new antenna element, named Walled-LTSA, and integrated the elements to construct an integrated antenna, with which we demonstrated a quick reflection measurement and a good performance to visualize plastic landmines. This year, we analyzed the effect of the wall shape on the radiation characteristics quantitatively. Based on the result, we optimized the wall morphology. Another work was the integration of the three functional modules, i.e., microwave, CSOM-classification, and landmine-class identification modules. We constructed an integrated handset. It attracted many researchers in conferences and exhibitions such as MWE 2005.
2 Interferometric Synthetic Aperture Radars
A.Hirose and R.Yamaki

We proposed a novel phase unwrapping method in which we spread the singularity in phase map with fractional phase compensators, unlike we do in the conventional network-flow cost minimization (NFCM) method compensating with discrete 2np. We demonstrated that obtained digital elevation map has higher quality than those obtained in the NFCM method. In addition, the calculation cost was found very small.

3 Highly-functional Adaptive Ultrasonic Imaging
A.Hirose and T.Nishino
Based on the theory of the complex-valued neural networks, we investigate highly-functional adaptive ultrasonic imaging. Conventional ultrasonic systems have wide room for improvement in imaging and shape-reconstructing performance. Phase-sensitive adaptive processing is making a big breakthrough.
4 Adaptive Dispersion Compensation in Optical Communications
A.Hirose and K.Tanizawa
In the near-future all-optical networks, adaptive dispersion compensation is indispensable for high-speed flexible communications. Last year, we proposed a low-cost dispersion-measurement-free compensation method that deals with time-domain signal-waveform directly and controls compensation device in frequency domain. We formulated the dynamics. This year, we demonstrated a fast compensation. with a VIPA dispersion compensator.
5 Fast 3D-Movie Generation by CGH Using Complex-Valued Neural Networks
A.Hirose, K.Tanizawa and T.Higo
Computer-generated hologram (CGH) requires a long calculation time. In particular, the generation of a CGH stream for three-dimensional movies takes a huge amount of calculation cost. We proposed a small-calculation-cost method to generate a CGH stream based on coherent neural networks that deal with complex-amplitude information with generalization ability in the frequency domain. We demonstrated successfully that, after carrier- frequency- dependent learning, we can generate a CGH stream by sweeping the carrier frequency with neural interpolation thanks to frequency-domain generalization.
6 Influence and Utilization of Delay in Neural Sensorimotor Systems
A.Hirose and Y.Azuma
We investigate the influence of delay in sensorimotor systems on motion control and its utilization. Humans have inevitable delay of approximately 100ms in sensing and motion. Therefore, we cannot perform a good motion control with a simple feedback mechanism. Some researchers say that, because of this delay, we have been able to develop our intelligence. We believe that the delay should also be effective in artificial neural networks. We investigate the mechanism with which neural networks utilize the delay in sensorimotor systems in learning and self-organization.
7 Extraction and Processing of Letters/Characters in Natural Images
A.Hirose and J.Koyama
We conducted experimental study on extraction and processing of letters/characters in natural images. In particular, we have developed a new method realized in Fourier domain. The detail is to be presented.
8 Quantitative Analysis of Neuron Behavior
A.Hirose, M.Watanabe and M.Morinaga
Instead of the conventional cable theory, we previously formulated a potential-based pulse propagation theory on the neuron membrane. Our new theory is free from parameter arbitrariness so that we can connect microscopic physics directly to mesoscopic neuron activity. This year, we applied it to (1)evaluation of pulse speed along axon with/without myelin and (2)analysis of spine shaping mechanism.
Publications List
[1] S.Kawata A.Hirose : “Frequency multiplexed logic circuit based on coherent optical neural network,” Applied Optics, 44, 19 (2005) 4053-4059.
[2] A.B.Suksmono A.Hirose: “Beamforming of ultra-wideband pulses by a complex-valued spatio-temporal multilayer neural network,” International Journal of Neural Systems, 15, 1-2 (2005) 85-91.
[3] T.Hara A.Hirose: “Adaptive plastic-landmine visualizing radar system: Effects of aperture synthesis and feature-vector dimension reduction,” IEICE Trans. on Electron., E88-C, 12 (2005) 2282-2288.
[4] A.B.Suksmono A.Hirose: “Progressive transform-based phase unwrapping utilizing a recursive structure,” IEICE Trans. on Commun. , E89-B, 3 (2006) 929-936.

[5] L.Amornrat S.Kawata A.Hirose: “Phase signal embedment in densely frequency-multiplexed coherent neural networks,” OSA Topical Meeting on Information Photonics (OSA_IP) 2005 Charlotte, Proc. (June 6-9, 2005, Charlotte) ITuA2.
[6] T.Oishi A.B.Suksmono A.Hirose: “Proposal of bilinear surface compensation of distortion in least-squrares phase unwrapping,” Int'l Geoscience and Remote Sensing Symposium (IGARSS) 2005 Seoul, Proc. (July 25-19, 2005, Seoul) 168.1.
[7] K.Tanizawa A.Hirose: “Gradient-Method Based Adaptive Control of Tunable Dispersion Compensator that Minimizes Time-Domain Waveform Error,” Optical Fiber Commun. Conf. (OFC), 2006 Los Angeles, Proc. (March 5-10, 2006, Los Angeles) CA, OWI-45.

[8] 谷澤 健, 廣瀬 明: “時間波形誤差を最小化する勾配法を用いた適応的分散補償,” 信学会光通信システム研究会, OCS2005-40 (2005).
[9] 増山 惣一, 廣瀬 明: “高速・高空間分解能の地中イメージング用集積アンテナ・ハンドセット,” 信学会/電気学会電磁界理論研究会, EMT05-82 (2005).
[10] 渡辺将也, 廣瀬明: “Nernst-Planck 方程式に基づく軸索活動電位伝搬速度の3次元解析,” 信学会ニューロコンピューティング研究会, NC2005-111 (2006-3).
[11] 廣瀬 明, 浅野 泰史, 濱野 敏彦: “コヒーレント型ニューラルネットワークにおける動作モードの修飾と発展学習,” 第15回インテリジェント・システム・シンポジウム, 2A2-4 (2005) 297-302.
[12] L.Amornrat, S.Kawata, A.Hirose: “Phase signal embedment in densely frequency-multiplexed coherent neural networks,” Frontiers in Optics 2006 Rochester, Proc. (October 8-12, 2005, Rochester) FTuY5 .
[13] A.Hirose: “Complex-valued self-organizing map for imaging antipersonnel plastic landmines: A brain brought up with phase-sensitive eyes,” Int'l Symposium on Electronics for Future Generations, COE21 --"Secure-life electronics" for quality life and society--, Proc. (October 11-12, 2005, Tokyo) 445-451.
[14] 増山惣一, 廣瀬 明: “高速・高空間分解能の地中イメージング用集積アンテナ・ハンドセット,” 第6回21世紀COEワークショップ, Proc. (March 13, 2006, Tokyo) 55-60.

[15] 廣瀬 明, 谷澤 健: “波形劣化補償方法及び装置,” 特許出願, 特願2005-163358.

[16] 廣瀬明, 電子情報通信学会エレクトロニクスソサイエティ功労賞, (社)電子情報通信学会, 25-Mar-06