Select Language

AI社区

公开数据集

触摸传感器事件

触摸传感器事件

3.73M
434 浏览
0 喜欢
0 次下载
0 条讨论
Earth and Nature,Computer Science,Programming,Classification,Neuroscience Classification

数据结构 ? 3.73M

    Data Structure ?

    * 以上分析是由系统提取分析形成的结果,具体实际数据为准。

    README.md

    Context A very simple dev board based on a esp32 was build. It reads 8 touch sensors arranged in a 2x4 array. The esp32 detects an inital touch on any sensor and then sends the (normalized) values of each touch sensor for the next 100 points (~one second ) to a couchdb database over a wifi connection. Due to the setup 8*100 values are recorded per event. The idea is to train i.e. 3 arbitrary gestures and use and ML to classifiy new events. The touch array is intended to work as a general userinterface. For the inital test, the following gestures have been used and the 'class' name is given. swipe right = "r" swipe left = "b" double tap = "k" As one can see in the overview of the three gesutes, even a human beeing would have no problem distunguishing the the profiles. The idea is not to write a algorithm that does the classification, but use a neuronal net to classify the 800 input features. This work is intended to get into ML and a POC, to show that a neuronal network is able to distingush pattern. Also the idea is to give an idea of an general jupyter notebook (python) based workflow example. ![enter image description here][1] ![enter image description here][2] Content The supplied dataset may be loaded into pandas and will result in a dataframe with a datetime bases index. This is the time, when the document was created on the couchdb. The "event" column holds the array with the touch sensor values. So there are 100x8 values. The numpy equivalent shape would be (100, 8). The "class" holds a string which was given when recording the events. As a consequence of the inital idea of a POC, the class have been referring to thier color in the later repsentation (red, blue and black). This was not changed yet is is gore sure not the best way to have a comprehensiv dataset. Acknowledgements Andrew Ng, for his work on coursera. And the micropython and esp-idf community for bringing python to the esp32. Inspiration Can a neuronal network take away the part of writing the algorithm? Is it enough if I supply the idea and the code is written by the network? [1]: https://storage.googleapis.com/kagglesdsdata/datasets/28422/36213/esp32_touch.JPG?GoogleAccessId=web-data@kaggle-161607.iam.gserviceaccount.com&Expires=1527504284&Signature=a3Be8ZTbBoU36luiwe%2BRolgeIiv9i2TgDDRlSPl3QbzUu2cSyOEqGNHiZBEuTik4FRtikx9RhHgBZo3iYIUuD6mmUufLmQogGRQgLs8vEAKRgwmfnvJ8jrDp1WvS67DcQiBwN1GtjCvmx%2Fx2feO%2F6eqC%2Fl1evo1tNrJzaV9KNfsvfT0DAsMeMltJqZ5jfozhPxHbHIekqMiCdlNwqg3yGTowb5Jpral3aat4wQOCf74j5iLr3Rmg%2FJ%2B0hV5sKonq%2F%2FvEtZ056HmW4cbx%2BIzM9qWGPptxuQBrBViWtrP4aULcL%2BNtwP4z9XaoBAfoqUmkOP9e9h4K6U6jQjI9%2BSvOpA%3D%3D [2]: https://storage.googleapis.com/kagglesdsdata/datasets/28422/36219/event_overview.jpg?GoogleAccessId=web-data@kaggle-161607.iam.gserviceaccount.com&Expires=1527514694&Signature=OIvLKRqWNLuBUBptUx7DFNKc6Pp1KPCB2TkdCwDPerrRmJ7WObqKUqYicFU3zhgYf13Aqi%2B7DZJ5vGLCxhvZWePoNh6fb%2FqOeISrsuwkZ8Aw66Mwbnl%2BHqttjskf9qGq9O7R2jFDOUWfrrss28%2FzgxkTUayftwx1E%2FdSy32TilKgfx1XQgsyeiPgsjUPMXkgMhw9Me%2B5fr4Oxq8QriUuL1Ft0MvZynsMdHPZtj49R25NND9E3HjWRmuRmXaz7ADucVWc4Ljwj7Crxp6pRAPwx0BmHoxzrBbbKcLoz355I%2BngWqZKEIMQfkZKwpGdVkBw9j%2FKBsgi%2BBLnDCjzl8T5bA%3D%3D
    ×

    帕依提提提温馨提示

    该数据集正在整理中,为您准备了其他渠道,请您使用

    注:部分数据正在处理中,未能直接提供下载,还请大家理解和支持。
    暂无相关内容。
    暂无相关内容。
    • 分享你的想法
    去分享你的想法~~

    全部内容

      欢迎交流分享
      开始分享您的观点和意见,和大家一起交流分享.
    所需积分:0 去赚积分?
    • 434浏览
    • 0下载
    • 0点赞
    • 收藏
    • 分享