Yevgeniy Vorobeychik

Title: Yevgeniy Vorobeychik: Innovator in Image Classification Defense

Introduction

Yevgeniy Vorobeychik is a notable inventor based in St. Louis, MO (US). He has made significant contributions to the field of image classification, particularly in developing systems that defend against physical attacks on such technologies. His innovative approach addresses the vulnerabilities of image classification systems, making them more robust against adversarial threats.

Latest Patents

Vorobeychik holds a patent for "Systems and methods for defending against physical attacks on image classification." This patent describes an image classification system that defends against physically realizable attacks. The system retrieves a training dataset of input images and generates an adversarial image based on one of the selected input images. The adversarial image is created by occluding a portion of the selected image with a predetermined shape, such as a rectangle, containing noise. A defense against occlusion attacks (DOA) classifier is then trained using the training dataset and the adversarial image. This classifier is utilized to classify captured images of items, such as street signs, that may have been subjected to attacks like sticker placement or vandalism. Vorobeychik's work in this area is crucial for enhancing the security of image classification systems.

Career Highlights

Yevgeniy Vorobeychik is affiliated with Washington University, where he continues to advance his research and innovations. His work has garnered attention for its practical applications in safeguarding image classification technologies. With a focus on developing effective defense mechanisms, Vorobeychik has established himself as a key figure in this domain.

Collaborations

Some of Vorobeychik's notable coworkers include Tong Wu and Liang Tong. Their collaborative efforts contribute to the ongoing research and development in the field of image classification and security.

Conclusion

Yevgeniy Vorobeychik's contributions to image classification defense highlight the importance of innovation in technology security. His patent and ongoing research efforts are paving the way for more secure image classification systems.