A New TV World for Kids - When ZUI Meets Deep Learning

In this work, we propose a novel application of Zoomable User Interface (ZUI) for TV world. With the utilization of latest advances of deep learning, visual tags of video titles can be semi-automatically extracted. By creating a cascaded visual tag tree structure, the ZUI representation problem is converted into an optimization problem of choosing tags to achieve the least user interactions. To the best of our knowledge, this is the world’s first effort in adopting ZUI on TV display with an optimized solution to achieve the best user experiences. Experimental results indicate that this invention is favored by kids, due to its nature of visual richness, intuitive, and user friendly.

[1]  Wei Liu,et al.  SSD: Single Shot MultiBox Detector , 2015, ECCV.

[2]  Ali Farhadi,et al.  You Only Look Once: Unified, Real-Time Object Detection , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[3]  Benjamin B. Bederson,et al.  A review of overview+detail, zooming, and focus+context interfaces , 2009, CSUR.

[4]  David J. Fleet,et al.  Efficient Non-greedy Optimization of Decision Trees , 2015, NIPS.

[5]  Zhihai He,et al.  Spatially supervised recurrent convolutional neural networks for visual object tracking , 2016, 2017 IEEE International Symposium on Circuits and Systems (ISCAS).

[6]  Benjamin B. Bederson.  The promise of zoomable user interfaces , 2010, VINCI '10.

[7]  Haohong Wang,et al.  VideoTopic: Modeling User Interests for Content-Based Video Recommendation , 2014, Int. J. Multim. Data Eng. Manag..