幼儿是祖国的花朵,是国家未来发展的希望,在幼儿的成长过程中,国家及家长重视教育方法,教育质量,教育环境,但是在关注幼儿的智力,身体健康的同时,也要重视幼儿心理健康的发展,在幼儿没有成熟的思想,极易被影响时,幼儿生活环境因素就尤...幼儿是祖国的花朵,是国家未来发展的希望,在幼儿的成长过程中,国家及家长重视教育方法,教育质量,教育环境,但是在关注幼儿的智力,身体健康的同时,也要重视幼儿心理健康的发展,在幼儿没有成熟的思想,极易被影响时,幼儿生活环境因素就尤为重要。Children are the flowers of our motherland and the hope for the future development of the country. In the process of children’s growth, the country and parents attach great importance to education methods, education quality, and education environment. However, while paying attention to children’s intelligence and physical health, they should also attach importance to the development of children’s psychological health. When children do not have mature thoughts and are easily influenced, the factors of their living environment are particularly important.展开更多
针对低层特征对图像内容描述不够精确而导致现场勘验图像(crime scene investigation,CSI)分类准确率低的不足,结合特征融合与几何短语池化提出了一种高效图像特征编码和融合方法。首先,分别提取图像密集SIFT和边缘SIFT特征并进行融合;...针对低层特征对图像内容描述不够精确而导致现场勘验图像(crime scene investigation,CSI)分类准确率低的不足,结合特征融合与几何短语池化提出了一种高效图像特征编码和融合方法。首先,分别提取图像密集SIFT和边缘SIFT特征并进行融合;然后,采用几何短语池化技术对融合特征进行编码,并利用多尺度空间金字塔匹配产生包含空间位置信息的稀疏编码特征;最后,通过迁移学习提取图像深度卷积特征,与编码后的特征融合成最终图像特征,并采用支持向量机对图像进行分类。实验结果表明,与经典的图像分类算法相比,所提方法更适合于现场勘验图像分类并取得了较高的分类准确率。展开更多
文摘幼儿是祖国的花朵,是国家未来发展的希望,在幼儿的成长过程中,国家及家长重视教育方法,教育质量,教育环境,但是在关注幼儿的智力,身体健康的同时,也要重视幼儿心理健康的发展,在幼儿没有成熟的思想,极易被影响时,幼儿生活环境因素就尤为重要。Children are the flowers of our motherland and the hope for the future development of the country. In the process of children’s growth, the country and parents attach great importance to education methods, education quality, and education environment. However, while paying attention to children’s intelligence and physical health, they should also attach importance to the development of children’s psychological health. When children do not have mature thoughts and are easily influenced, the factors of their living environment are particularly important.
文摘针对低层特征对图像内容描述不够精确而导致现场勘验图像(crime scene investigation,CSI)分类准确率低的不足,结合特征融合与几何短语池化提出了一种高效图像特征编码和融合方法。首先,分别提取图像密集SIFT和边缘SIFT特征并进行融合;然后,采用几何短语池化技术对融合特征进行编码,并利用多尺度空间金字塔匹配产生包含空间位置信息的稀疏编码特征;最后,通过迁移学习提取图像深度卷积特征,与编码后的特征融合成最终图像特征,并采用支持向量机对图像进行分类。实验结果表明,与经典的图像分类算法相比,所提方法更适合于现场勘验图像分类并取得了较高的分类准确率。
