本文選題：視覺(jué)注意 + 自頂向下��； 參考：《華中科技大學(xué)》2016年碩士論文

【摘要】：人在看一副圖像時(shí),會(huì)不自覺(jué)的關(guān)注圖像中某些區(qū)域,同時(shí)忽略某些區(qū)域。這種視覺(jué)感知過(guò)程中表現(xiàn)出的選擇性是視覺(jué)注意機(jī)制作用的結(jié)果。在計(jì)算機(jī)視覺(jué)研究中,通過(guò)對(duì)視覺(jué)注意機(jī)制進(jìn)行建模,可以賦予計(jì)算機(jī)在復(fù)雜環(huán)境中自動(dòng)獲取人的視覺(jué)興趣區(qū)的能力。通過(guò)模擬人類視覺(jué)感知系統(tǒng),研究人員提出了基于特征整合的視覺(jué)注意計(jì)算框架。在此框架下衍生出了多種視覺(jué)注意模型。本文詳細(xì)分析Itti和Judd兩種具有較大影響力的顯著性視覺(jué)注意模型。其中Itti模型通過(guò)整合多種低層次特征生成顯著圖作為圖像區(qū)域受關(guān)注程度的預(yù)測(cè)。該方法忽略視覺(jué)注意過(guò)程中知識(shí)、任務(wù)、偏好等因素的影響。Judd模型整合高層次語(yǔ)義特征作為知識(shí)的引入方式。雖然取得了較好的效果,但是啟發(fā)式特征的設(shè)計(jì)和計(jì)算較復(fù)雜,擴(kuò)展性不強(qiáng)。本文在現(xiàn)有視覺(jué)注意模型基礎(chǔ)上重點(diǎn)研究了兩個(gè)問(wèn)題:(1)如何通過(guò)學(xué)習(xí)方法獲取視覺(jué)注意特征。(2)如何在特征整合框架下進(jìn)行層次特征整合。首先,本文通過(guò)訓(xùn)練卷積神經(jīng)網(wǎng)獲取像素級(jí)、對(duì)象級(jí)、語(yǔ)義級(jí)特征。然后,基于學(xué)習(xí)獲取的特征,提出了一種整合層次特征的視覺(jué)注意模型,重點(diǎn)在于利用對(duì)象屬性信息進(jìn)行高層次特征整合,該方法有效彌補(bǔ)了已有模型在引入先驗(yàn)知識(shí)方面的不足。最后,針對(duì)提出的視覺(jué)注意模型,設(shè)計(jì)了一種層次知識(shí)引導(dǎo)的注意焦點(diǎn)轉(zhuǎn)移方法。實(shí)驗(yàn)表明,新模型充分利用了先驗(yàn)知識(shí),在多個(gè)數(shù)據(jù)集上測(cè)試均獲得了較好的實(shí)驗(yàn)結(jié)果。
[Abstract]:When you look at an image, you will unconsciously focus on some areas of the image, while ignoring some areas. The selectivity of visual perception is the result of visual attention mechanism. In the research of computer vision, by modeling the visual attention mechanism, the computer can automatically acquire the region of visual interest in complex environment. By simulating human visual perception systems, researchers proposed a visual attention computing framework based on feature integration. Under this framework, several visual attention models are derived. Two significant visual attention models, Itti and Judd, are analyzed in detail. The Itti model uses a variety of low-level features to generate salient maps as a prediction of the attention level of the image region. This method ignores the influence of knowledge, task, preference and other factors in visual attention. Judd model integrates high-level semantic features as a way to introduce knowledge. Although good results have been obtained, the design and calculation of heuristic features are more complicated and less extensible. Based on the existing visual attention models, this paper focuses on two problems: (1) how to acquire visual attention features through learning methods; (2) how to integrate hierarchical features in the framework of feature integration. Firstly, the features of pixel level, object level and semantic level are obtained by training convolution neural network. Then, based on the features acquired by learning, a visual attention model integrating hierarchical features is proposed, which focuses on the high-level feature integration using object attribute information. This method effectively makes up for the deficiency of the existing models in introducing prior knowledge. Finally, aiming at the proposed visual attention model, a method of attention focus shift based on hierarchical knowledge guidance is designed. Experiments show that the new model makes full use of prior knowledge, and good experimental results are obtained by testing on multiple data sets.
【學(xué)位授予單位】：華中科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前2條

1 暴林超;蔡超;肖潔;周成平;;一種用于復(fù)雜目標(biāo)感知的視覺(jué)注意模型[J];計(jì)算機(jī)工程;2011年13期

2 肖潔;蔡超;丁明躍;;一種圖斑特征引導(dǎo)的感知分組視覺(jué)注意模型[J];航空學(xué)報(bào);2010年11期

，

本文編號(hào)：2052293