摘要
传统监督哈希方法将图像学习的手工特征或机器学习特征和二进制码的单独量化步骤分开,并未很好地控制量化误差,并且不能保证生成哈希码的平衡性。为了解决这个问题,提出了新的多尺度平衡深度哈希方法。该方法采用多尺度输入,这样做有效地提升了网络对图像特征的学习效果;提出了新的损失函数,在很好地保留语义相似性的前提下,考虑了量化误差以及哈希码平衡性,以生成更优质的哈希码。该方法在CIFAR-10以及Flickr数据集上的最佳检索结果较当今先进方法分别提高了5. 5%和3. 1%的检索精度。
The use of the semantic similarity improving the hash coding quality has recently been more widely concerned. Traditional supervised hash methods for image retrieval represent an image as a manual feature vector or a machine learning feature vector,and then perform a separate quantization step to generate a binary code. Such methods do not control the quantization error effectively,and cannot guarantee the balance of hash code. To this end,this paper presented a new multi-scale balanced deep hash method. The method used multi-scale input,which effectively improved the ability of learning the image features from the network. Moreover,it proposed a new loss function. Under the premise of preserving the semantic similarity,it took the quantization error and the balance of hash code into account to generate the high quality hash code. After experimenting on two benchmark databases: CIFAR-10 and Flickr,this method has been improved by 5. 5% and 3. 1% of the search accuracy compared with today's advanced image retrieval methods.
引文
[1]易唐唐,黄立宏. CBIR中一种基于最近邻的改进相关反馈算法[J].计算机应用研究,2015,32(8):2326-2330.(Yi Tangtang,Huang Lihong. Improved relevance feedback algorithm based on nearest neighbor approach in CBIR[J]. Application Research of Computers,2015,32(8):2326-2330.)
[2] Raginsky M,Lazebnik S. Locality-sensitive binary codes from shift-invariant kernels[C]//Advances in Neural Information Processing Systems.[S. l.]:NIPS,2009:1509-1517.
[3] Norouzi M,Blei D M. Minimal loss hashing for compact binary codes[C]//Proc of International Conference on Machine Learning. Bellevue:Omnipress,2011:353-360.
[4] Gong Yunchao,Lazebnik S,Gordo A,et al. Iterative quantization:a procrustean approach to learning binary codes for large-scale image retrieval[J]. IEEE Trans on Pattern Analysis and Machine Intelligence,2013,35(12):2916-2929.
[5] Weiss Y,Torralba A,Fergus R. Spectral hashing[C]//Advances in Neural Information Processing Systems. Vancouver:Curran Associates Inc.,2009:1753-1760.
[6] Wang Jun,Kumar S,Chang S F. Semi-supervised hashing for scalable image retrieval[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Piscataway,NJ:IEEE Press,2010:3424-3431.
[7] Strecha C,Bronstein A,Bronstein M,et al. LDAHash:improved matching with smaller descriptors[J]. IEEE Trans on Pattern Analysis and Machine Intelligence,2012,34(1):66-78.
[8] Liu Wei,Wang Jun,Ji Rongrong,et al. Supervised hashing with kernels[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Piscataway,NJ:IEEE Press,2012:2074-2081.
[9] Zhang Peichao,Zhang Wei,Li Wujun,et al. Supervised hashing with latent factor models[C]//Proc of the 37th International ACM SIGIR Conference on Research&Development in Information Retrieval. New York:ACM Press,2014:173-182.
[10] Krizhevsky A,Sutskever I,Hinton G E. Imagenet classification with deep convolutional neural networks[J]. Communications of the ACM,2017,60(6):84-90.
[11] Xia Rongkai,Pan Yan,Lai Hanjiang,et al. Supervised hashing for image retrieval via image representation learning[C]//Proc of the28th AAAI Conference on Artificial Intelligence. Palo Alto,CA:AAAI Press,2014:2156-2162.
[12] Lai Hanjiang,Pan Yan,Ye Liu,et al. Simultaneous feature learning and hash coding with deep neural networks[C]//Proc of International Conference on Computer Vision and Pattern Recognition. Piscataway,NJ:IEEE Press,2015:3270-3278.
[13]Lazebnik S,Schmid C,Ponce J. Beyond bags of features:spatial pyramid matching for recognizing natural scene categories[C]//Proc of International Conference on Computer Vision and Pattern Recognition.Piscataway,NJ:IEEEPress,2006:2169-2178.
[14] He Kaiming,Zhang Xiangyu,Ren Shaoqing,et al. Spatial pyramid pooling in deep convolutional networks for visual recognition[J].IEEE Trans on Pattern Analysis and Machine Intelligence,2015,37(9):1904-1916.
[15]Hao Jiedong,Dong Jing,Wang Wei,et al. What is the best practice for CNNs applied to visual instance retrieval?[EB/OL].(2016-11-05). https://arxiv. org/abs/1611. 01640.
[16] Girshick R. Fast R-CNN[C]//Proc of International Conference on Computer Vision and Pattern Recognition. Piscataway,NJ:IEEE Press,2015:1440-1448.
[17] Tolias G,Sicre R,JéGou H. Particular object retrieval with integral max-pooling of CNN activations[EB/OL].(2016-02-24). https://arxiv. org/abs/1511. 05879.
[18]Wang Jingdong,Shen Hengtao,Song Jingkuan,et al. Hashing for similarity search:a survey[EB/OL].(2014-08-13). https://arxiv. org/abs/1408. 2927.
[19] Li Wujun,Wang Sheng,Kang Wangcheng. Feature learning based deep supervised hashing with pairwise labels[EB/OL].(2016-04-21). https://arxiv. org/abs/1511. 03855.
[20] Kang Wangcheng,Li Wujun,Zhou Zhihua. Column sampling based discrete supervised hashing[C]//Proc of the 30th AAAI Conference on Artificial Intelligence. Palo Alto,CA:AAAI Press,2016:1230-1236.
[21]Do T T,Doan A Z,Cheung N M. Discrete hashing with deep neural network[EB/OL].(2015-08-28). https://arxiv. org/abs/1508.07148.
[22]Zhu Han,Long Mingsheng,Wang Jianmin,et al. Deep hashing network for efficient similarity retrieval[C]//Proc of the 30th AAAI Conference on Artificial Intelligence. Phoenix:AAAI Press,2016:2415-2421.
[23]Kulis B,Darrell T. Learning to hash with binary reconstructive embeddings[C]//Advances in Neural Information Processing Systems. Vancouver:Curran Associates Inc.,2009:1042-1050.
[24] Gong Yunchao,Lazebnik S,Gordo A,et al. Iterative quantization:a procrustean approach to learning binary codes for large-scale image retrieval[J]. IEEE Trans on Pattern Analysis and Machine Intelligence,2013,35(12):2916-2929.