PornSeerTest Interface

After we push the button “Detect Image” from the PornSeer Pro interface, the porn image detection interface shown as follow will pop up. The functions of each portion of this interface are listed as follows.

1. Push “Open File” to open the image folder that you want to scan.

You most likely will see the following folder browser interface. Choose any file under the folder that you want to scan, in this example, the folder we want to scan is “..\inocent”

After you press the “Open” button, PronSeerTest will scan the image folder you just chose. This will take a while if there are many images under this folder and if you shift the focus out the PornSeerTest interface, this interface will be halt while it scans the folder. You don’t need to do anything but wait for the scanning ends.

2. After the scan finished, you can find the results from two newly generated folders under the image folder you chose; in our case, the folder “..\inocent”. PornSeerTest will put the detected porn image file under folder “..\inocent\porn” and the clean image under folder “..\inocent\none”. The following image shows a portion of the folder “..\inocent\none”.

Observe that in this folder that file names are followed a special convention: [degree to be porn]_[original file name]. For example, the file name “05.37_innocent (119).jpg” denotes that the original image file “innocent (119).jpg” has a degree to be porn of “05.37”. Similarly, the file named “00.00_innocent (177).jpg” denotes that the image file “innocent (177).jpg” has a degree to be porn of “00.00”. By using this name convention, we can easily to check the detection results generated by PornSeerTest and based on which we can refine the parameters that will be discussed next.

3. When the “Sensitive (More findings)” radio button is checked, we can find more results with big hit rate and also more false alarm.

4. When the “Accurate (Less findings)” radio button is checked, we can find less results with less hit rate and also less false alarm. Therefore, the results are more accurate because less false results are included.

5. The parameter of “Decision Threshold” can be set by the users as a cut-off point for porn and clear images. It is the threshold for the “degree to be porn” as shown in the image name convention just mentioned. The user can test a few sets of results to choose the optimal parameters for the specific application. The default parameters are the optimal value for Yang’s Benchmark Image Database for Porn-Detection Software.

The academic users can ask for a low-priced the academic developers’ kit and free source codes in C\C++ for non-commercial purpose. For commercial applications, the license fee depends on the functions needed. For more information on this technology, please send your request to