互動(dòng)視窗

        圖像、聲音等媒體識(shí)別技術(shù)的發(fā)展與應(yīng)用正在悄然地給人們生活帶來變革?；?dòng)視窗針對(duì)博物館文物展示等類似應(yīng)用場(chǎng)合，以對(duì)參觀者手勢(shì)、意念行為、語音等信息的識(shí)別和處理為技術(shù)基礎(chǔ)，設(shè)計(jì)一個(gè)方便、自然和人性化的嵌入式互動(dòng)平臺(tái)。

        作品在EC5-1719CLDNA上，針對(duì)展品和參觀者行為擴(kuò)展了攝像頭、拾聲器、可控展品臺(tái)等硬件裝置；研究了手勢(shì)識(shí)別與跟蹤、人臉檢測(cè)與跟蹤、非特定人中英文連續(xù)語音識(shí)別等可用于人機(jī)交互的新方法和算法；研究并改進(jìn)了基于手形結(jié)構(gòu)特征的識(shí)別算法，建立了手勢(shì)缺陷圖的判斷方法；在Linux上實(shí)現(xiàn)了具有物品展示、多媒體評(píng)論、娛樂等功能的互動(dòng)平臺(tái)。

        作品還針對(duì)平臺(tái)雙核的特點(diǎn)，對(duì)識(shí)別算法進(jìn)行并行化處理。圖像識(shí)別使用多線程的方式，對(duì)手勢(shì)識(shí)別、人臉檢測(cè)進(jìn)行并行處理；手勢(shì)識(shí)別的手勢(shì)分割和識(shí)別過程采用了OpenMP優(yōu)化計(jì)算，有效地提高了系統(tǒng)的運(yùn)行速度。

        全息物品展示功能可以理解參觀者意圖，讓參觀者立體而真實(shí)的觀看所展物品，隨心所欲的觀看每個(gè)細(xì)節(jié)；多媒體評(píng)論功能獲取參觀者的音頻和繪制的圖像，為參觀者提供了交流平臺(tái)；對(duì)展示物品的手勢(shì)拼圖游戲豐富了物品展示的功能，增加了娛樂性。

        互動(dòng)視窗中所實(shí)現(xiàn)的基于音、視頻識(shí)別的多種人機(jī)交互方式還可以有更廣闊的應(yīng)用。

The development and application of image, voice and other media recognition technology are quietly to bring changes to people's lives. The production is designed for museum and some similar places. We have designed a convenient, natural and customized embedded platform based on the recognition and analyze of the gesture, behavior and voice of the visitors.
Camera, microphone, controllable platform and other hardware are added to EC5-1719CLDNA. Gesture identification and tracking, face detection and tracking, human-independent English and Chinese continuous speech recognition have been researched; improved the recognition algorithm of hand-shaped structure, gesture concave map judgments; achieved features like display of items, multimedia reviews, entertainment on Linux.

To the characteristics of dual-core platform, we optimized the recognition algorithm for parallel processing. The gesture recognition, face detection are paralleled; the gesture division and identification process of gesture recognition are optimized by OpenMP, which effectively improved the speed of system.

Holographic display feature is capable to understand the visitor’s idea and show what they want to watch; multimedia feature achieve comment on the voice and drawing, providing a platform for communication; Puzzle Game that play by hand gesture enriched the system and added entertainment.

The multi-channel interaction which implemented by image and voice recognition is able to be used in more application.

linux操作系統(tǒng)文章專題:linux操作系統(tǒng)詳解（linux不再難懂）