May I ask which dataset are you training on?

[protossw512]

Thank you for providing pretrained model. I have a question regarding to the trainnig dataset. Did you trained on the dataset from original paper or something else?

I tried to do testing on doll.png and troll.png from http://www.alphamatting.com dataset and got pretty bad results, not sure what happened.

I modified test.py code so that I am directly using trimap provided by the dataset, instead of generating trimap from alpha channel.

The segmentation seems to be heavily based on white area on alpha channel.

Here is my result of troll.png:

I am new to image matting, not sure where I did wrong, do we need to do some preprocessing?

[Joker316701882]

@protossw512 I trained on company's dataset which can be partially seen in test_data folder, it's relatively simple. Dataset mentioned in paper is more complex so the model training on it can generalize very well. Unfortunately, I tried to train on dataset mentioned in paper, but it's really really hard to converge. I'll try to improve the performance on it in future. Thank you for your great comments!

[protossw512]

@Joker316701882 Thanks for reply, since I am working on university research project, I think I could get the dataset from author. I will update if I could get a better model.

[Joker316701882]

@protossw512 Looking forward to your result! Let me know if you make it!

[bhack]

@protossw512 If you have any progress send us an update.

[hieudt-pixelz]

In the paper, the authors mentioned this: "We find images on simple or plain backgrounds (Fig. 2a), including the 27 training images from [25] and every fifth frame from the videos from [26]"

Do you know how to get that video from [26]? I glanced over that paper but don't know which video to use. "[26] E. Shahrian, B. Price, S. Cohen, and D. Rajan. Temporally coherent and spatially accurate video matting. In Proceedings of Eurographics, 2012. 3"

[Ru-Xiang]

@protossw512 if you have had a progress, may you share the trained-model with us ?

[protossw512]

Yeah, I actually slightly modified loss function and data loading part, fixed some issue with original code that may lead to not good results, and also added refinement stage, getting some promising results, but I am not sure if I am allowed to release pretrained model, since I am training on Adobe dataset and I was not supposed to share it without permission of author. I haven't got time to refactor my code and post it on github yet. The model itself is not very hard to train, barely need to tune any parameters.

[Ru-Xiang]

Thank you for your reply, it will be very appreciated if you share your codes at your convenience.

[protossw512]

Yeah, I think I am going to do that during winter break

[Devin147]

@protossw512 Hi, can you tell me how to fix the code in detail except for refinement? Specially, what's the problem about loss function and network architecture?

[Joker316701882]

@protossw512 Glad to here that you achieved promising results! Can you email me your email address? If it's possible, I would like to discuss about what refinement you have made to achieve good results. Thank : )

[protossw512]

@Joker316701882 @ChengJunqi @Ru-Xiang Yeah, sure, please email me: helloxinyao.wang###gmail.com I think one fatal mistake for your original code is the creation of trimap from alpha, you have to perfom eroding and dilation alpha, but you only applied dilation, that's why the network is basically remember the boundaries of your trimap where pixel values are equal to 255.

There should be other minor issues, I cant remember them tho. I would like to rewrite your code but I do not have any more time to do so, but I can clean them a little bit and upload to my github

[ricshaw]

@protossw512 Did you manage to get a better model?

[foamliu]

@protossw512 Could you share the improvement in your github?

[elzeru]

Hello, if you can't have the dataset from Adobe, check on pixabay.com. There is a large amount of quite large pictures with transparency and good alpha. For instance, try this search : https://pixabay.com/fr/photos/?min_width=640&min_height=640&colors=transparent&image_type=photo&order=latest

The only issue is that you have to download the images by hand, but with a good methodology you can obtain 300 foreground image in a couple of hour.

Combined with the backgrounds described in the DIM publications, you can obtain encouraging results without any change to the network.

[shliang0603]

@protossw512 Can you send me the Adobe datasets, I have send a email for author of Deep Image Matting, but I didn't get a reply.Thanks

[18191171661]

@protossw512，Can you send me the Adobe datasets, I have send a email for author of Deep Image Matting, but I didn't get a reply.Thanks. My email is 1575262785@qq.com.

[tsing90]

@protossw512 I know you may not be able to share your trained model as you explained authorization problem, but would you please share Joker's pretrained model? the link of his pretrained is missing ...... I'll appreciate much if you could do me this favor, thanks. my email: LC2015@ic.ac.uk

[dailenson]

@protossw512，Can you send me the Adobe datasets, I have send a email for author of Deep Image Matting, but I didn't get a reply.Thanks. My email is 1575262785@qq.com.

have you got the dataset? thanks!

[jayhhl]

@protossw512 Can you send me the Adobe datasets, I have send a email for author of Deep Image Matting, but I didn't get a reply.Thanks. My email is 332907213@qq.com.

[ZhangChaoZhong]

Yeah, I think I am going to do that during winter break

Can you send me the Adobe datasets.Thanks!My email is 1014295391@qq.com

[ofrimasad]

can you please send me the Adobe dataset. The author is not answering emails. Thanks in advance. ofrimasad@yahoo.com

[dudumeng]

Could you send me the Adobe datasets? Thanks! My email is yeqingbit@qq.com

[linpengzhang]

@protossw512 Can you send me the Adobe datasets? Thanks! My email is linpeng.zhang@studenti.unipd.it I'm a CS Student and I need it for MSc thesis

[sinfere]

Could you send me the datasets? I'm doing some related researches. Thanks!

My email is sinfere@gmail.com

[flyfish01]

@protossw512

Could you share the Adobe datasets, thank you! my email is haoran.andy###gmail.com

[parachute0001]

@protossw512 Could you send me the datasets? I'm doing some related researches. Thanks! My email is 785003851@qq.com

[tengda100]

已经好久不研究这方面的技术了

------------------ 原始邮件 ------------------ 发件人: "parachute0001"notifications@github.com; 发送时间: 2019年10月22日(星期二) 上午9:58 收件人: "Joker316701882/Deep-Image-Matting"Deep-Image-Matting@noreply.github.com; 抄送: "Subscribed"subscribed@noreply.github.com; 主题: Re: [Joker316701882/Deep-Image-Matting] May I ask which dataset areyou training on? (#8)

@protossw512 Could you send me the datasets? I'm doing some related researches. Thanks! My email is 785003851@qq.com

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[confident-OH]

Thank you for providing pretrained model. I have a question regarding to the trainnig dataset. Did you trained on the dataset from original paper or something else?

I tried to do testing on doll.png and troll.png from http://www.alphamatting.com dataset and got pretty bad results, not sure what happened.

I modified test.py code so that I am directly using trimap provided by the dataset, instead of generating trimap from alpha channel.

The segmentation seems to be heavily based on white area on alpha channel.

Here is my result of troll.png:

I am new to image matting, not sure where I did wrong, do we need to do some preprocessing?

Could you share the Adobe datasets, thank you very much! My email is confidentoh@foxmail.com

[confident-OH]

@protossw512 Can you send me the Adobe datasets, I have send a email for author of Deep Image Matting, but I didn't get a reply.Thanks

Could you share the Adobe datasets, thank you very much! My email is confidentoh@foxmail.com

[confident-OH]

@protossw512 Can you send me the Adobe datasets? Thanks! My email is linpeng.zhang@studenti.unipd.it I'm a CS Student and I need it for MSc thesis

Could you share the Adobe datasets, thank you very much! My email is confidentoh@foxmail.com

[imocat]

@protossw512 could you send me the adobe dataset? thanks! my email is imocat###qq.com

[PytaichukBohdan]

@protossw512 I would appreciate very much if you could send me the Adobe datasets :) It is needed very much for my Masters Degree. My email is: bohdan.pytaichuk@gmail.com

[drnemor]

@protossw512 Could you send me Adobe dataset? Thank you for your work! My email: drnemor@gmail.com

[mingqizhang]

I apply the dataset from paper author, but did not get the respond. Could you please send me the Adobe datasets, my school mail is 161060075@hdu.edu.cn

[jaintarun]

@protossw512 Hi can you send me the Adobe dataset also? I emailed bprice@adobe.com but I have received no reply. I am doing a project in Image Matting and it will help a lot if I can get the dataset. My email address is tjain73@yahoo.com

@Joker316701882 @ChengJunQi @Ru-Xiang Yeah, sure, please email me: helloxinyao.wang###gmail.com I think one fatal mistake for your original code is the creation of trimap from alpha, you have to perfom eroding and dilation alpha, but you only applied dilation, that's why the network is basically remember the boundaries of your trimap where pixel values are equal to 255.

There should be other minor issues, I cant remember them tho. I would like to rewrite your code but I do not have any more time to do so, but I can clean them a little bit and upload to my github

[SvenGroen]

Has anyone access to the dataset and is able to share it with me. I am working on my bachelor thesis and desperately looking for a benchmark dataset! Feel free to send it to me: sgroen@uni-osnabrueck.de

[cisspbru]

Hello, could you please send me the Adobe dataset. The author is not answering emails. I'm a spbgpu student and I needed it very much my email 6032083@mail.ru

[alexey-trushkov]

@protossw512

Hi, would you mind sharing with me Adobe dataset. I'm doing some related research. My email: alexeitrushkov@gmail.com.

[luckyboy101]

Can you send me the Adobe datasets.Thanks!My email is 2783000@qq.com

[ghost]

@protossw512 Hi! Could you possible share with me Adobe datasets. My email is pseudomo@yandex.ru Thanks!

[luckyboy101]

Hi, I am not allowed to share this dataset due to the adobe license agreement. Please contact Brian Price (bprice@adobe.com), and you will need to sign an agreement.

------------------ 原始邮件 ------------------ 发件人: "Joker316701882/Deep-Image-Matting" <notifications@github.com>; 发送时间: 2020年10月18日(星期天) 上午6:53 收件人: "Joker316701882/Deep-Image-Matting"<Deep-Image-Matting@noreply.github.com>; 抄送: "loske"<2783000@qq.com>;"Comment"<comment@noreply.github.com>; 主题: Re: [Joker316701882/Deep-Image-Matting] May I ask which dataset are you training on? (#8)

@protossw512 Hi! Could you possible share with me Adobe datasets. My email is pseudomo@yandex.ru Thanks!

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[qingyang-jiandao]

@protossw512 Can you send me the Adobe datasets, I have send a email for author of Deep Image Matting, but I didn't get a reply.Thanks. My email is 695450439@qq.com.

[zpy-number]

Hi, would you mind sharing with me Adobe dataset. I'm doing some related research. My email:17302632110@163.com

[luckyboy101]

Hi, I am not allowed to share this dataset due to the adobe license agreement. Please contact Brian Price (bprice@adobe.com), and you will need to sign an agreement.

------------------ 原始邮件 ------------------ 发件人: "zpy-number"<notifications@github.com>; 发送时间: 2021年3月5日(星期五) 凌晨0:35 收件人: "Joker316701882/Deep-Image-Matting"<Deep-Image-Matting@noreply.github.com>; 抄送: "loske"<2783000@qq.com>; "Comment"<comment@noreply.github.com>; 主题: Re: [Joker316701882/Deep-Image-Matting] May I ask which dataset are you training on? (#8)

Hi, would you mind sharing with me Adobe dataset. I'm doing some related research. My email:17302632110@163.com

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[cnnAndBn]

hi,@Joker316701882,I have a question about the dataset. in paper ,the fig 2,b shows the alpha result of fig 2 a by using ps tools right? but fig2 c seems to not correspond well to fig 2 c. as the following image show, the alpha image is very accurate but the corresponding image in fg folder is not good as highlighted in the blue circle. what is the images in the fg folder? how they are generated? thks

[heran]

@protossw512 Could you share the Adobe datasets, thank you very much! My email is wenqiang#xmdqq.com

[Gundoganfa]

Could you share the Adobe datasets, thank you very much! My email is gundoganfa@gmail.com

[AloneGu]

@protossw512 Could you send me Adobe Matting dataset? Thank in advance! My email: wg631@nyu.edu

[luckyboy101]

Hi, I am not allowed to share this dataset due to the adobe license agreement. Please contact Brian Price @.***), and you will need to sign an agreement.

------------------ 原始邮件 ------------------ 发件人: "Jackling @.>; 发送时间: 2021年6月18日(星期五) 晚上6:04 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [Joker316701882/Deep-Image-Matting] May I ask which dataset are you training on? (#8)

@protossw512 Could you send me Adobe Matting dataset? Thank in advance! My email: @.***

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.