摘要
针对盲人群体在超市购物存在视觉障碍的问题,笔者设计一款基于深度学习的盲人辅助购物APP。设计的功能包括商品种类识别、商品颜色识别、商品标签识别以及语音播报。该APP采用轻量级神经网络MobileNet和单阶段目标检测算法SSD完成商品种类识别,借助Tesseract-OCR技术完成商品标签识别,利用讯飞语音SDK完成语音播报,拥有广阔的应用前景。
Aiming at the visual impairment of blind people shopping in supermarkets,the author designed an assisted shopping APP for blind people based on deep learning.The designed functions include product type recognition,product color recognition,product label recognition,and voice broadcast.The APP uses lightweight neural network MobileNet and single-stage target detection algorithm SSD to complete product category recognition,Tesseract-OCR technology to complete product label recognition,and iFLYTEK Voice SDK to complete voice broadcast.It has broad application prospects.
作者
张洁
刘芬芬
谢宏
邵秋丹
林建
ZHANG Jie;LIU Fenfen;XIE Hong;SHAO Qiudan;LIN Jian(College of Computer Science and Technology,Guizhou University,Guiyang Guizhou 550025,China)
出处
《信息与电脑》
2021年第4期103-104,共2页
Information & Computer
基金
贵州省高等学校大学生创新创业训练计划项目“基于人工智能的盲人辅助购物APP的设计与实现”(项目编号:202010657047)
贵州大学高等教育研究项目“本科导师制人才创新能力培养模式研究”(项目编号:Gdgj2016008)。
