Since traditional image feature extraction methods rely on features such as corner points, a new method based on semantic feature extraction is proposed inspiring by convolution neural attack. The semantic features of each pixel in an image are computed and quantified by neural network to represent the contribution of each pixel to the image semantics. According to the quantization results, the semantic contribution values of each pixel are sorted, and the semantic feature points are selected from high to low and the image mosaic is completed. Experimental results show that this method can effectively extract image features and complete image mosaic.

中文翻译：

---

基于卷积神经网络语义特征提取的图像拼接方法

由于传统的图像特征提取方法依赖于角点等特征，因此提出了一种基于卷积神经攻击的基于语义特征提取的新方法。通过神经网络计算和量化图像中每个像素的语义特征，以表示每个像素对图像语义的贡献。根据量化结果，对每个像素的语义贡献值进行排序，并从高到低选​​择语义特征点，完成图像拼接。实验结果表明，该方法可以有效地提取图像特征并完成图像拼接。