zkisthebest / siamese-rpn Goto Github PK

In _get_64_anchor(self,gtbox), can you explain why the x of anchor is 14+28a in the 'anchor = [14+28a, 14+28*b, self.anchor_shape[c][0], self.anchor_shape[c][1]]', I can not understand since the whole network stride is 15. @zkisthebest

So sorry to interrupt u.I just occured an error when i trained SiamRPN," gtbox = os.listdir(self.root_dir + folder + '/label/')[index]" was always alerted.Maybe it is caused by the former of annotation files,i guess.
All the groundtruth was written in a file groundtruth.txt.I really am grateful if u could offer me some suggestions!!!

[69] img = os.listdir(self.root_dir + folder + '/img/')[index]
[74]gtbox = os.listdir(self.root_dir + folder + '/label/')[index]
这两行代码，取得的图片和真值并不对应，listdir（）遍历的文件名可能是乱序的，因此并不对应。
还有测试代码里也是。。。

Hello,
In your donwload.py file you use two files, namely detection_train.txt and detection_train.txt. However, these files are not the annotation files from the official Youtube-BB dataset. I am wondering where to download this two files?

Many thanks,
Yiming

Thanks for your sharing. We found this version cant backward properly, so we have re-implement it.
pytorch-version is available here: [https://github.com/songdejia/siamese-RPN-pytorch]
tensorflow-version is available here: [https://github.com/makalo/Siamese-RPN-tensorflow.git]
I hope this can help you.

Excuse me, I would be very grateful if you could provide the pretrained model what you trained by the OTB data or youtube-bb data.

感谢你复现了这个算法。请问有训练好的模型，效果如何呢？