基于多尺度感受野的警觉保持计算模型-计算机软件与理论专业论文.docxVIP

摘爱摘要 摘爱 摘要 计算枧视觉是人工智熊研究领域中一个十分活跃的分支，其研究领域中存在 的一个重要难点是效率、精度与计算资源间的平衡问题。生物视觉系统将注意汇 聚予首要售惑款曩时，还对周透信愚绦持警觉。生甥程甍系统孛存在弱这季孛诗算 资源分配机制有效地缓解了巨量计算的负担。因此，模仿生物视觉系统建立相应 静计算模型箴为解决讦葬辊视觉中存在的诸多难题韵一个有效手段。 在生物视觉系统中，视赏信息处理煦第一阶段发生在视网膜，视觉信息经视 网膜后，已被进行了初步的处理。因此，视网膜的输出可被理解为对图象迸一步 分辑兹黔颈处理。本文通过模数人类褪鹚膜榜经节绥照如僖息处理枧裁，获褥多 尺度的感受野分布，设计并实现了在对视野中心区域感兴趣的信息进行处理的同 时，对瘸边僖怠傈撩～定警觉的艨次瘸络诗算穰鍪，为实魏计算撬程觉系统豹实 时陛和避免巨量计算提供肖价值的启示。 本文提出的层次网络计算模疆以高斯差模型(difference ofGaussian，DOG)为 具体计冀手段，通过On型与Off型DOG算予作用在多尺度感受野的反应竣出， 检测视野图象内发生的不同尺度的On事件与Off事件，以判断视野图象是否存 在显蓍交纯，扶嚣决定是否转移程点，逐孪亍更遴一步麓精缨处理。滚计算模壅邋 过仿真实验取得了良好的实验结果。 关键词：计算机视觉、神经视觉、图象理解、高鼓羞模型、感受野 分类号：TP391．4 AbstractComputer Abstract Computer vision is an active branch of Artificial Intelligence．An important problem in this area is the balance among efficiency,accuracy and computing resource．Human visual system can keep watchfulness to the perimeter of visual field while their central attention is focused to the center of visual field for fine information processing．This mechanism of computing resource assignment could ease the demand for huge and complex hardware structure．Therefore designing computer model based olt biological visual mechanism is an effective approach to resolve problems in computer vision Retina is the first step of visual information processing in biological visual system，Visual information is processed preliminarily after retina processing．Therefore，the output of retina can be considered霸s the pre堙eatment for further information processing．In this paper a multi—layer neural model was designed based on the features of receptive field of ganglion in retina to simulate multi-scale receptive fields of ganglion cell．The model could maintain the alert to the outer area of the whole image while it was capturing and processing more important information locating at the image’s central part．It may provide vaIuabte inspiration for the implementation of real—time processing and the avoidance ofhuge computing in computer vision． The computation model proposed in this the