Журавлев Алексей Юрьевич AbsoluteGZ

Hello, World!

😎 About Me :

I am a Python Developer from Russia.

🔭 I'm learning Python, which is a powerful and flexible programming language that's ideal for getting started and progressing in development, but I don't intend to stop there.
🌱 I implement projects using the Pytorch, TensorFlow and Keras libraries to better understand the principles of operation of neural networks.
⚡ I regularly read articles and articles on forums such as Stack Overflow and Habr to broaden my horizons.
😈Started publishing his projects on GitHub, demonstrating the use of Python and neural networks to solve real-world problems.
🧠Very interested in building interactive apps and websites using Python and machine learning to personalize user experience.
💯I aim to achieve key knowledge and skills that will allow me to implement powerful and innovative projects in the future.

My significant projects :

1) Internship CV Engineer at Wildberries - https://github.com/AbsoluteGZ/WBTECH_Learn

EX2_Quality

Photos uploaded by WB suppliers have different quality: Some may have a complex background, some photos may have part of the object not in the frame, etc. To work with such data later, for example, when using photo search algorithms, you need to know the type of defects/features present in the image. There are 6 types of features in total:

untidy, angle-composition, background, crop, text, multiple-objects and one class for images without defects good-image. In this case, the image may contain several types of defects.

Task:

It is necessary to create a model that will determine the list of defects for the image.

EX3_Fraud

Users upload images to the reviews section. Unscrupulous users use reviews to advertise their products or services. As a rule, such images contain contact information: phone numbers, a link to a website, a telegram account, etc.

Required:

Develop an algorithm/software system that allows distinguishing acceptable photos from unacceptable ones. The difficulty is due to the fact that acceptable photos can also contain text.