You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
open_nsfw_android/README.md

2.5 KiB

open_nsfw_android

Mozilla Add-on

色情图片离线识别,基于TensorFlow实现。识别只需200ms,可断网测试,成功率99%,调用只要一行代码,从雅虎的开源项目open_nsfw_python移植,tflite(6M)为训练好的模型,该模型文件可用于iOS、java、C++等平台

使用

  • Add it in your root build.gradle at the end of repositories:
	allprojects {
		repositories {
			...
			maven { url 'https://jitpack.io' }
		}
	}
  • Add the dependency
	dependencies {
	         //versionCode:上面小icon中最新版本号
	        implementation 'com.github.devzwy:open_nsfw_android:[versionCode]'
	}

  • Code like this
   val nsfwBean = NsfwHelper.getInstance(this, true, 1).scanBitmapSyn(bitmap)
   nsfwBean?.sfw ...
   nsfwBean?.nsfw ...

Demo运行结果:

图片

0.9254902 0.7764706 0.7254902 0.92156863 0.77254903 0.72156864

[16:16+224,16:16+224,:]       [16:240,16:240,:]

0.90588236-R 0.75686276-G 0.68235296-B ------0.01960784 0.01960784 0.04313724 0.90196079 0.75294119 0.6784314

[:,:,:: -1]

0.68235296-B 0.75686276-G 0.90588236-R

*255 转换为int

174 193 231

-[104, 117, 123]

70 76 108

opencv假设图像是RGB三分量组成的图像,那么图像的

第一通道是R,

第二通道是G,

第三通道是B

Img[:,:,2]代表R通道,也就是红色分量图像;

Img[:,:,1]代表G通道,也就是绿色分量图像;

Img[:,:,0]代表B通道,也就是蓝色分量图像。