Introduction and short description

Table of contents

Module Creation Studio

This tool allows you to create your own modules for captcha recognition, which can already be used in CapMonster2.

The technologies used in the new version of CapMonster2 are very different from the previous version, so the training steps may seem complicated and incomprehensible at first glance. In these articles, we will try to explain everything in as much detail as possible, everything in text form and more simply, using examples in the video. But, if you still have questions after reading, please ask us on the forum, we will help and be sure to supplement these articles based on your feedback. So, let's begin.

You can download the latest version of CapMonster2 MCS x64 here.

Creating your own recognition module

Creation of your own recognition module is divided into several steps.

Here we will provide only a brief description of each step and a link to a more detailed description.

If something is not clear, it is better to ask on the forum in the program section. Otherwise, you may end up wasting your time on useless work.

Project creation

You need to open the program, create a new project, save it under a name that you understand.

Then add captchas to the project, for which you need to create a recognition module. Save.

To work with these captchas further, they must be recognized and, preferably, with 100% reliability.

Stage 1. Creation of a project and collection of captchas.

Collecting symbols

Recognition of captchas is, in fact, recognition of the characters that are written on it, i.e. it is necessary to train the module to recognize characters. And to teach the module to recognize characters, these characters need to be collected.

The collection is carried out manually: you need to click with the mouse in the center of the symbol.

Stage 2. Collecting symbols.

Setting up filters

You can significantly increase the recognition by applying conventional graphic filters to the captcha.

For example, you can increase the contrast or highlight a single color.

The most useful thing is to enlarge the captcha with small characters for better recognition and reduce the captcha with large characters for faster module operation.

Stage 3. Filters.

Setting up the search for centers of mass

As mentioned earlier, captcha recognition is the recognition of characters on it, but before you can recognize a character, you need to find it. It is necessary to search for a symbol in special points - potential centers of symbols.

In the same place, the size of the character recognition window is selected - a rectangular area in which the largest representative of characters will fit.

Stage 4. Centers of mass.

Module training

When all the resources are prepared and the basic settings are made, you need to start training the module.

First, you need to adjust the complexity of the core depending on how badly the characters are distorted. A complex core is a large percentage of recognition over a long time of work, and vice versa, a simple core is a module with a lower percentage of recognition, but it works much faster.

After configuring the complexity of the core, you need to configure the training process itself.

The faster the training, the worse the quality of the received module.

Training is the longest part of creating a module and can take a day in especially difficult cases.

Stage 5. Module training.

Recognition test and module quality improvement

After training, you need to test the finished module.

During training, it will also be constantly tested and the percentage of recognition will be visible immediately during training. The point of this step is to select simple but important parameters and check whether the recognition percentage of a ready-made module will improve at the same time.

Also, at this stage you can find the main causes of recognition errors and try to eliminate them.

Stage 6. Testing and improvement.

Importing a ready-made recognition module

Module Creation Studio only trains modules and stores them together with captchas and other information not necessary for direct recognition of captchas. Therefore, after the module has been trained, it must be saved in an appropriate format that CapMonster2 understands. It is it who will receive and recognize captchas instead of manual recognition services.

When the work on creating the module is completed and the percentage of recognition you are satisfied, you need to click on "Save" in the "Testing module" mode and specify the path where your own ready-made module will be saved. As a result, you will receive a ready-made recognition module, which you can then add to CapMonster2.

Stage 7. Saving.