ReCAPTCHA Dataset

As the title said, I will be giving away thousands of reCAPTCHA images dataset, since I don’t use it anymore. The dataset is also used in my previous project.

The dataset consists of the most popular 3×3 reCAPTCHA. For example, bicycle, bridge, bus, car, crosswalk (or pedestrian crossing for you British people), hydrant, mountain or hill, palm, etc. The dataset includes the “Other” folder, which consists of images that are not classified, or random objects.

Before I wrote this post, I am planning to sell it, but then I changed my mind and decided to gave it away for free. So, please consider donating as it took hundreds of hours for me to collect it.

Preview of the dataset:

The images are compressed and available for download in the links below.

I suggest you download the sample images first and decide if the dataset is suitable for you, before downloading the larger dataset.

Sample (80 images, 3 MB): Sample Download

Large (11.000 images, 400 MB): Large Download

Since Google Drive hosts the files, please comment if the links is not working anymore.

Project Ideas

Feel free to create anything with the dataset. Here are some project ideas using the dataset.

• reCAPTCHA solver (I built this before, check this post).
• Generate new CAPTCHA images.
• Image classification, mainly for vehicle-related (cars, bus, traffic light, crosswalk, etc) images.

Thanks for reading. I hope you can build a success project. If so, please comment below what project have you made using this reCAPTCHA dataset.

from Hacker News https://ift.tt/2BxFX8v