Deep Learning for Design Conditioned by Brain Signals

项目介绍 Project Introduction

人们一直以来都认为读取思想只存在于科幻小说中而不是科学现实,所以在这一领域研究跨出的每一小步都是一大飞跃。 事实上,在神经科学领域,关于由脑电图和功能磁共振成像捕获的脑活动信号以及如何处理此类信号的研究,已经成为一个非常活跃且相对成功的研究领域。 这种关于脑活动信号的分析可以进一步应用于设计。 人们观察、思考和决策的方式都可能反映在产品设计或广告中。对于生产商来说,为客户设计和选择最具吸引力和最相关的图像是一项巨大的挑战。

通过使用深度神经网络模型,我们此次研究课题已经能够将脑信号分类,并达到80.56%的准确率,信号分类被用于制约图像生成过程。

Reading mind has been regarded to more of a science fiction than real science. Each little step in this area will make a big difference. Research on Neuroscience on EEG and fMRI captured brain activity signals and how to process these signals have been a very active and relatively successful field. This analysis of brain activity signals can be further applied to design. The way people observe, think, and make decisions may all be reflected in the design of a product or an advertising. Designing and choosing the most engaging and relevant images for their client is a big challenge.

Using deep neural network models, our study theme has been able to classify brain signals up to an accuracy of 80.56%, which were used to picture the image generation process.

团队介绍 Team Introduction

导师 Mentor

吴超浙江大学公共管理学院研究员 WU Chao, Research professor

导师助理 Assistants

王小毅 副教授 WANG Xiaoyi, Associate professor

王盼博士生 WANG Pan, Doctoral Candidate

李骏翔 博士生 LI Junxiang, Doctoral Candidate

潘虹安 硕士 PAN Hongan, Master

参与学生 Students

Ashlyn Goh Er Xuan – 吴而萱, Edric, Lim Keng Hin – 林敬轩, Ng Au Ker Wesson, Wan Zhi Jun – 万芷君, Yee Celine- 余思霖, Yong Khai Sheen – 杨凯勋

支持公司 Supporting Company

中国通信服务浙江工程公司

China Comservice Zhejiang Construction Co., Ltd

theme2_label_poster

RELEVANT INFORMATION