- Python
- Opencv
- tensorflow (if using nn_play.py)
- Adb tools
- Android Phone
for IOS (Refer to this site for installation)
- iPhone
- Mac
- WebDriverAgent
- facebook-wda
- imobiledevice
- multiscale-search
- CV based fast-search
- Convolutional Neural Network based coarse-to-fine model
Notice: CV based fast-search only support Android for now
It is recommended to run the following if have an android phone
python play.py --phone Android --sensitivity 2.045
If you have an iPhone, download the model following the link bolow, and run the following
python nn_play.py --phone IOS --sensitivity 2.045
--phonehas two options: Android or IOS.--sensitivityis the constant parameter that controls the pressing time.play.pyusing algorithm based on CV, support Android and IOSnn_play.pyusing algorithm based on Convolutional Neural Network, support Android and IOS, recommend for IOS
Our method can correctly detect the positions of the man (green dot) and the destination (red dot).
It is easy to reach the state of art as long as you like. But I choose to go die after 859 jumps for about 1.5 hours.
Here is a video demo. Excited!
CNN train log and train&validation data avaliable at
Training: you need to download and untar data directory and change data dir path in data_provider
Inference: you need to download and unzip train log dirs and put them in dir named resource