cnn生成图像显著图_一种利用DCGAN提高基于CNN图像识别性能的方法技术

本发明专利技术公开了一种利用DCGAN提高基于CNN图像识别性能的方法，该方法将DCGAN出色的数据生成能力与基于CNN图像识别框架进行了二度结合，并且DCGAN是在GAN的基础上经过改进后的新型对抗生成网络，所述方法将CNN应用到了原始结构中，使得GAN具有了深度卷积的特性，并在数据生成方面拥有更好的特征表示形式。本发明专利技术很好的解决了图像识别过程中训练样本数据难以收集、样本相似度过大等问题，冲破了样本数量与质量在分类模型优化问题上的限制，进一步强化分类模型，提高图像识别的准确性。

A method of improving the performance of CNN based image recognition using DCGAN

The invention discloses a method for improving CNN image recognition performance by using DCGAN. The method combines the excellent data generation ability of DCGAN with the CNN-based image recognition framework, and DCGAN is a new type of antagonistic generation network improved on the basis of GAN. The method applies CNN to the original structure. In this way, GAN has the characteristics of deep convolution, and has a better feature representation in data generation. The invention solves the problems of difficult collection of training sample data and excessive similarity of sample in the process of image recognition, breaks through the limitation of sample quantity and quality in the optimization of classification model, further strengthens classification model and improves the accuracy of image recognition.

【技术实现步骤摘要】

一种利用DCGAN提高基于CNN图像识别性