基于深度卷积神经网络和局部敏感哈希的图像检索

网络图像资源增长迅速,如何实现快速有效的大规模图像检索,成为当前研究的热点之一。深度神经网络对图片特征有很强的表达能力,利用典型深度卷积神经网络VGG16在预训练完成的模型上使用网络全连接层的输出提取待检索图像数据集的特征以建立索引,并采用局部敏感哈希算法提升检索速度,以端到端的形式,完成基于内容的图片检索任务。这种图像检索模型提供了一种在计算资源有限情况下实现大规模图像检索的有效方法。

Image Retrieval Based on Deep Convolutional Neural Network and Locality Sensitive Hash

Network image resources are growing rapidly. How to achieve fast and effective large-scale image retrieval has become one of the hotspots of current research. In this paper, the typical deep convolutional neural network VGG16 is used to extract the features of the image dataset to be retrieved using the output of the network connection layer on the pre-trained model to establish the index, and the local sensitive hash algorithm is used to improve the retrieval speed and performs an end-to-end content-based image retrieval tasks. The image retrieval method designed in this paper provides an effective method for large-scale image retrieval under limited computing resources.