About LSTM model

[EricPengShuai]

Where is the LSTM part in your code? After reading it, I think the following function is quite similar to the function of the LSTM part, right？

LiveDeep/VideoProcessingPMultipleParameters.py

Line 115 in 8d291c7

def UpdateTileStatUF(TileStaByUF,UWL, UHL ,UWH, UHH):

def UpdateTileStatUF(TileStaByUF,UWL, UHL ,UWH, UHH):
    for i in range(len(TileStaByUF)):
        TileStaByUF[i]=0
    iL = int(math.floor(UWL / 256))
    jL = int(math.floor(UHL / 144))
    iH = int(math.floor(UWH / 256))
    jH = int(math.floor(UHH / 144))
    Flag = 0
    if iH >= 5:
        iH = 4
    if jH >= 5:
        jH = 4
    for i in range(iL,iH+1):
        for j in range(jL,jH+1):
            TileStaByUF[j * 5 + i]=1

[XianglongFeng]

To follow up and clarify: First, I am very sorry for the late response. I was graduating in the past few months and did not catch up with the GitHub issues posted in this repo. Second, I checked the version of the code in the repo and realized that it was not the newest version and did not contain the LSTM code you were asking. I have now updated the code, and the LSTM part is in “estUserTrace()” in “VideoPip.py”. I apologize for the inconvenience. Please let me know if you have any further questions.

[EricPengShuai]

@XianglongFeng Excuse me, could please tell me how calculate viewer's eye coordinates according to quaternions referring to the following line of your code?

LiveDeep/VideoPip.py

Line 236 in 69d60c9

H, W = LocationCalculate(x, y, z, w)

[XianglongFeng]

Please check the reference [41]. Thank you very much.

[lieDoeCherif]

I'm now doing research with LiveDeep, but I find that the code on github is not complete.
Can you send me a full version of the code ?mail: chorica@tutanota.com. Thanks a lot.

[XianglongFeng]

I'm now doing research with LiveDeep, but I find that the code on github is not complete. Can you send me a full version of the code ?mail: chorica@tutanota.com. Thanks a lot.

We should have released all the LiveDeep code. Can you please let us know what code you found missing in the GitHub repo?

[lieDoeCherif]

In"VideoProcessingPMultipleParameters.py", Line25: " from misc import progress_bar", is there any file named "misc"? I don't find it.

Line 27, what's the definition of MyOwnModel ?

In function "CheckPredictionResult", Line77, FlagB has no definition.

[XianglongFeng]

In"VideoProcessingPMultipleParameters.py", Line25: " from misc import progress_bar", is there any file named "misc"? I don't find it.

Line 27, what's the definition of MyOwnModel ?

In function "CheckPredictionResult", Line77, FlagB has no definition.

1. Please check:
   https://www.freshports.org/misc/py-progressbar
   for the details about the progress_bar. You can also install it using pip.

2. "MyOwnModel " is used to test customized models. If you use the existing model, you can just use the models in the folder "models ". If not, you can ignore this and the "main" class.

3. Sorry for this typo, the "FlagB" is the same as the "FLagA" in the next function. You can check the updated one.