1. Download the CAPTCHA from specified websites
2. Segment and recognize the distorted and sticky characters from picture
Technologies:
1. Use Java multi-thread and HTTP protocol to download CAPTCHA
2. Use Otsu threshold to binarize the original image
3. Use CFS (Color Filling Segmentation) algorithm to implement the first segmentation
4. Use BIDF (Big Inertial Drop Fall) algorithm to implement the second segmentation
5. Use LibSvm to build the model and predict the class of characters
请发表评论