为了减小训练集中各类别资源分布不均衡对分类性能造成的影响,该文对原始训练集使用类别均衡法,即对原始训练集以类为单位进行重新组合,使得重组后的训练集类别分布尽可能均衡,从而可以在均衡的类别上进行训练和分类,以降低在训练过程...为了减小训练集中各类别资源分布不均衡对分类性能造成的影响,该文对原始训练集使用类别均衡法,即对原始训练集以类为单位进行重新组合,使得重组后的训练集类别分布尽可能均衡,从而可以在均衡的类别上进行训练和分类,以降低在训练过程中对小类别的不公平待遇。在复旦大学语料库上使用类别均衡法,分别用N a ve B ayes和R occh io方法分类,前者的宏平均F1从48.62%提高到了80.99%,后者的宏平均F1从64.58%提高到80.26%,微平均F1从73.99%提高到80.47%。实验结果显示,类别均衡法显著提高了分类性能。展开更多
Generating photo-realistic images from a text description is a challenging problem in computer vision.Previous works have shown promising performance to generate synthetic images conditional on text by Generative Adve...Generating photo-realistic images from a text description is a challenging problem in computer vision.Previous works have shown promising performance to generate synthetic images conditional on text by Generative Adversarial Networks(GANs).In this paper,we focus on the category-consistent and relativistic diverse constraints to optimize the diversity of synthetic images.Based on those constraints,a category-consistent and relativistic diverse conditional GAN(CRD-CGAN)is proposed to synthesize K photo-realistic images simultaneously.We use the attention loss and diversity loss to improve the sensitivity of the GAN to word attention and noises.Then,we employ the relativistic conditional loss to estimate the probability of relatively real or fake for synthetic images,which can improve the performance of basic conditional loss.Finally,we introduce a category-consistent loss to alleviate the over-category issues between K synthetic images.We evaluate our approach using the Caltech-UCSD Birds-200-2011,Oxford 102 flower and MS COCO 2014 datasets,and the extensive experiments demonstrate superiority of the proposed method in comparison with state-of-the-art methods in terms of photorealistic and diversity of the generated synthetic images.展开更多
文摘为了减小训练集中各类别资源分布不均衡对分类性能造成的影响,该文对原始训练集使用类别均衡法,即对原始训练集以类为单位进行重新组合,使得重组后的训练集类别分布尽可能均衡,从而可以在均衡的类别上进行训练和分类,以降低在训练过程中对小类别的不公平待遇。在复旦大学语料库上使用类别均衡法,分别用N a ve B ayes和R occh io方法分类,前者的宏平均F1从48.62%提高到了80.99%,后者的宏平均F1从64.58%提高到80.26%,微平均F1从73.99%提高到80.47%。实验结果显示,类别均衡法显著提高了分类性能。
基金supported by the National Natural Science Foundation of China(Grant Nos.61972298 and 61962019)by the National Cultural and Tourism Science and Technology Innovation Project(2021064)the Training Program of High Level Scientific Research Achievements of Hubei Minzu University under Grant PY22011.
文摘Generating photo-realistic images from a text description is a challenging problem in computer vision.Previous works have shown promising performance to generate synthetic images conditional on text by Generative Adversarial Networks(GANs).In this paper,we focus on the category-consistent and relativistic diverse constraints to optimize the diversity of synthetic images.Based on those constraints,a category-consistent and relativistic diverse conditional GAN(CRD-CGAN)is proposed to synthesize K photo-realistic images simultaneously.We use the attention loss and diversity loss to improve the sensitivity of the GAN to word attention and noises.Then,we employ the relativistic conditional loss to estimate the probability of relatively real or fake for synthetic images,which can improve the performance of basic conditional loss.Finally,we introduce a category-consistent loss to alleviate the over-category issues between K synthetic images.We evaluate our approach using the Caltech-UCSD Birds-200-2011,Oxford 102 flower and MS COCO 2014 datasets,and the extensive experiments demonstrate superiority of the proposed method in comparison with state-of-the-art methods in terms of photorealistic and diversity of the generated synthetic images.
