At the CVPR conference, the prestigious test-of-time award was presented, which is given for an outstanding contribution to the development of computer vision. Each year, one publication is selected that was published exactly ten years earlier and had a significant impact on the field. This award is called the Longuet-Higgins Prize.
This year’s laureate was the renowned paper “Going Deeper with Convolutions,” which introduced the GoogLeNet architecture for the first time. This work became a breakthrough, opening new horizons in the creation of complex neural network models.
Interestingly, as early as 2014, this model won the ImageNet competition. It was one of the first examples of deep neural networks, demonstrating how to increase the number of layers without the risk of parameter explosion, while maintaining high model efficiency.
Congratulations on this well-deserved award!