In the talk, we will present the alarming results from our measurement study that most main-stream object detection models are adopting those winning image classification models in the ImageNet contest as their first few layers, to extract low-level features in the image.
By Yunhan Jia, Zhenyu Zhong, Yulong Zhang, Qian Feng, Tao Wei and Yantao Lu
Full Abstract & Presentation Materials: https://www.blackhat.com/asia-19/briefings/schedule/#the-cost-of-learning-from-the-best-how-prior-knowledge-weakens-the-security-of-deep-neural-networks-13723
