captcha_solver icon indicating copy to clipboard operation
captcha_solver copied to clipboard

Source code for ACM CCS 2018

Results 8 captcha_solver issues
Sort by recently updated
recently updated
newest added

非常感谢作者,读了论文,感觉作者的提出的想法很赞,通过合成验证码来实现小样本的验证码识别。有一个问题想要请教,论文提到先用generator生成验证码,然后利用GAN网络实现验证码的像素级的微调,让合成验证码和真实的更像。我想知道这个微调的GAN网络结构是怎样的?能具体解释下吗?期待得到您的回复 个人分析此处GAN网络应该是Image-based conditional GAN,但要实现像素级的微调,感觉有点困难,希望得到作者指点

Q1:can your solver solve the scene that there are repetitive characters in an image Q2:can the solver output the characters same as the marshalling sequence in the real image 问题1:验证码求解器可以应对验证码中有的字符出现多次的情况吗...

首先感谢作者的这篇论文,给了很多启发。 认真的读了论文,有几处不太明白的地方,希望能解答。 ## 我读论文理解的为 - captcha generator生成器生成两种验证码(with security and without security),这个生成器是Opencv或者Pillow生成的而不是模型 - captcha synthesizer是通过输入with security(上一步生成)输出为长得像real-like captcha,判别器模型输出captcha和真实captcha - preprocessing是通过输入captcha synthesizer的生成结果生成 without security图片(第一步生成的),判别器区分模型输出结果和第一步的without security captcha > 疑问:但是看Figure11和12,经过preprocessing又不像是生成的without security,所以preprocessing的目标图片怎么得来的 - 后面的LeNet的输入时without security captcha(经过preprocessing)...

Hi I have used the same parameters to train the LeNet model, but it is giving very high training loss and the loss keeps on increasing exponentially with the first...

Your paper used 200,000 synthetic images to train the LeNet5 model. I used the same number of data sets, but my accuracy and loss are declining, and I think there...

Can please share me the 500 sample dataset for synthetic image generator , that would be greatly helpful for my studies .

Do you have the plan to share the captcha preprocessing codes? I read your paper, very nice.

您好,在读您的论文时,感觉在Synthesizer部分有一些前后矛盾的地方。 在fiqure4中,您指出captcha image generator只负责在文字部分添加security features,而背景噪声这些则是交给网络来生成。但是在figure5里,您又指出类似于背景/斜线/噪声需要参数指定。那么背景噪声到底是由网络学习出来的,还是由第一步captcha image generator就已经生成出来了?