Closed tongxiaozhong14 closed 4 years ago
xuebinqin
commented
4 years ago Thanks for your interests. Currently, different size input may trigger some errors because the upsampling operations with factor of 2. To handle arbitrary input size, you can change the upsample functions as what we did in our new model U^2-Net:
def _upsample_like(src,tar): src = F.upsample(src,size=tar.shape[2:],mode='bilinear') return src
tongxiaozhong14
commented
4 years ago 您好，非常感谢你的回答，但是我有一个疑问？如果我拿u2net去测试视频的话，需要再针对数据集训练吗？还是说用别的已经训练好的数据集就可以实现视频检测发自我的华为手机-------- 原始邮件 --------发件人： Xuebin Qin notifications@github.com日期： 2020年6月18日周四 03:08收件人： NathanUA/BASNet BASNet@noreply.github.com抄送： tongxiaozhong14 tongxiaozhong14@163.com, Author author@noreply.github.com主 题： Re: [NathanUA/BASNet] The size of the input image required by the model training its own data (#39) Closed #39.
—You are receiving this because you authored the thread.Reply to this email directly, view it on GitHub, or unsubscribe.
tongxiaozhong14
commented
4 years ago 您好，我的意思是，训练出来的同一个模型同时用来做视频和图像的显著性检测，效果是一样的吗？非常感谢您的回答，你的回答将对我非常有帮助发自我的华为手机-------- 原始邮件 --------发件人： Xuebin Qin notifications@github.com日期： 2020年6月18日周四 03:08收件人： NathanUA/BASNet BASNet@noreply.github.com抄送： tongxiaozhong14 tongxiaozhong14@163.com, Author author@noreply.github.com主 题： Re: [NathanUA/BASNet] The size of the input image required by the model training its own data (#39) Closed #39.
—You are receiving this because you authored the thread.Reply to this email directly, view it on GitHub, or unsubscribe.
xuebinqin
commented
4 years ago 效果取决于视频帧与训练图像的相似性，我们在DAVIS测试了，如果视频中有显著性物体，效果还可以，但是某些帧之间的预测会出现不连续的情况。
On Sun, Jun 21, 2020 at 5:58 AM tongxiaozhong14 notifications@github.com wrote:
您好，我的意思是，训练出来的同一个模型同时用来做视频和图像的显著性检测，效果是一样的吗？非常感谢您的回答，你的回答将对我非常有帮助发自我的华为手机-------- 原始邮件 --------发件人： Xuebin Qin notifications@github.com日期： 2020年6月18日周四 03:08收件人： NathanUA/BASNet BASNet@noreply.github.com抄送： tongxiaozhong14 < tongxiaozhong14@163.com>, Author author@noreply.github.com主 题： Re: [NathanUA/BASNet] The size of the input image required by the model training its own data (#39) Closed #39.
—You are receiving this because you authored the thread.Reply to this email directly, view it on GitHub, or unsubscribe.
— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/NathanUA/BASNet/issues/39#issuecomment-647118328, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADSGORJ2PAUXE2LLQ2ZZ26TRXXYURANCNFSM4OAQKJMQ .
-- Xuebin Qin PhD Department of Computing Science University of Alberta, Edmonton, AB, Canada Homepage:https://webdocs.cs.ualberta.ca/~xuebin/
tongxiaozhong14
commented
4 years ago 感谢你的回答，现在有一个问题请教你一下，如果想要制作自己的数据集的话，有什么标注数据集的工具推荐吗？还是PS效果会好一些？发自我的华为手机-------- 原始邮件 --------发件人： Xuebin Qin notifications@github.com日期： 2020年6月22日周一 03:25收件人： NathanUA/BASNet BASNet@noreply.github.com抄送： tongxiaozhong14 tongxiaozhong14@163.com, Author author@noreply.github.com主 题： Re: [NathanUA/BASNet] The size of the input image required by the model training its own data (#39) 效果取决于视频帧与训练图像的相似性，我们在DAVIS测试了，如果视频中有显著性物体，效果还可以，但是某些帧之间的预测会出现不连续的情况。
On Sun, Jun 21, 2020 at 5:58 AM tongxiaozhong14 notifications@github.com
wrote:
您好，我的意思是，训练出来的同一个模型同时用来做视频和图像的显著性检测，效果是一样的吗？非常感谢您的回答，你的回答将对我非常有帮助发自我的华为手机--------
原始邮件 --------发件人： Xuebin Qin notifications@github.com日期： 2020年6月18日周四
03:08收件人： NathanUA/BASNet BASNet@noreply.github.com抄送： tongxiaozhong14 <
tongxiaozhong14@163.com>, Author author@noreply.github.com主 题： Re:
[NathanUA/BASNet] The size of the input image required by the model
training its own data (#39)
Closed #39.
—You are receiving this because you authored the thread.Reply to this
email directly, view it on GitHub, or unsubscribe.
—
You are receiving this because you modified the open/close state.
Reply to this email directly, view it on GitHub
https://github.com/NathanUA/BASNet/issues/39#issuecomment-647118328, or
unsubscribe
https://github.com/notifications/unsubscribe-auth/ADSGORJ2PAUXE2LLQ2ZZ26TRXXYURANCNFSM4OAQKJMQ
.
--
Xuebin Qin
PhD
Department of Computing Science
University of Alberta, Edmonton, AB, Canada
Homepage:https://webdocs.cs.ualberta.ca/~xuebin/
—You are receiving this because you authored the thread.Reply to this email directly, view it on GitHub, or unsubscribe.
xuebinqin
commented
4 years ago LabelMe, Matlab自带的标注工具箱，Photoshop, ByLabel (ubuntu only), etc.
On Thu, Jul 2, 2020 at 10:22 AM tongxiaozhong14 notifications@github.com wrote:
感谢你的回答，现在有一个问题请教你一下，如果想要制作自己的数据集的话，有什么标注数据集的工具推荐吗？还是PS效果会好一些？发自我的华为手机-------- 原始邮件 --------发件人： Xuebin Qin notifications@github.com日期： 2020年6月22日周一 03:25收件人： NathanUA/BASNet BASNet@noreply.github.com抄送： tongxiaozhong14 < tongxiaozhong14@163.com>, Author author@noreply.github.com主 题： Re: [NathanUA/BASNet] The size of the input image required by the model training its own data (#39) 效果取决于视频帧与训练图像的相似性，我们在DAVIS测试了，如果视频中有显著性物体，效果还可以，但是某些帧之间的预测会出现不连续的情况。
On Sun, Jun 21, 2020 at 5:58 AM tongxiaozhong14 notifications@github.com
wrote:
您好，我的意思是，训练出来的同一个模型同时用来做视频和图像的显著性检测，效果是一样的吗？非常感谢您的回答，你的回答将对我非常有帮助发自我的华为手机--------
原始邮件 --------发件人： Xuebin Qin notifications@github.com日期： 2020年6月18日周四
03:08收件人： NathanUA/BASNet BASNet@noreply.github.com抄送： tongxiaozhong14 <
tongxiaozhong14@163.com>, Author author@noreply.github.com主 题： Re:
[NathanUA/BASNet] The size of the input image required by the model
training its own data (#39)
Closed #39.
—You are receiving this because you authored the thread.Reply to this
email directly, view it on GitHub, or unsubscribe.
—
You are receiving this because you modified the open/close state.
Reply to this email directly, view it on GitHub
https://github.com/NathanUA/BASNet/issues/39#issuecomment-647118328, or
unsubscribe
< https://github.com/notifications/unsubscribe-auth/ADSGORJ2PAUXE2LLQ2ZZ26TRXXYURANCNFSM4OAQKJMQ>
.
--
Xuebin Qin
PhD
Department of Computing Science
University of Alberta, Edmonton, AB, Canada
Homepage:https://webdocs.cs.ualberta.ca/~xuebin/
—You are receiving this because you authored the thread.Reply to this email directly, view it on GitHub, or unsubscribe.
— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/NathanUA/BASNet/issues/39#issuecomment-653103597, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADSGORNV53TC5DJNQKSVA3LRZSX35ANCNFSM4OAQKJMQ .
-- Xuebin Qin PhD Department of Computing Science University of Alberta, Edmonton, AB, Canada Homepage:https://webdocs.cs.ualberta.ca/~xuebin/
Dear Author: Thank you for sharing the code. If I want to use your code to train my own data, is there any requirement for the size of the input training image?