图像检索(花卉)

分享一个用基于Pytorch和Flask实现的花卉图像检索系统(Flowers-Retrieval)，算法采用的是深度哈希，距离度量的方式利用的是汉明距离。以下是检索系统的界面：

 

 

 

以下是实现该系统的匹配代码的关键方法： 

    def match(self, uri, top_k=100):
        print(uri)
        uri = uri.replace('\\', '/')
        f = open('Data/search.txt', 'w')
        f.write(uri + ' 0')
        f.close()
        searchloader = load_data('Data/search.txt')
        search_binary = binary_output(searchloader)
        search_binary = search_binary.cpu().numpy()
        search_binary = np.asarray(search_binary, np.int32)
        print(search_binary)
        # tensor([[0., 1., 1., 0., 1., 0., 1., 0., 0., 1., 0., 0., 1., 1., 1., 0., 1., 1.,
        #          0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 0., 1., 0., 1.,
        #          1., 1., 0., 1., 1., 0., 0., 0., 1., 1., 1., 1.]])
        # 这里计算与train_binary最相似的一组,查询数据库
        query_result = np.count_nonzero(search_binary != self.binary, axis=1)  # don't need to divide binary length
        sort_indices = np.argsort(query_result)
        imlist = [self.imNames[index] for i, index in enumerate(sort_indices[0:top_k])]
        rank_dists = [query_result[index] for i, index in enumerate(sort_indices[0:top_k])]
        # print(imlist)
        print(rank_dists)
        for i in range(len(imlist)):
            # 获取文件路径后把把反斜杠替换为正斜杠
            imlist[i]='Data/Raw/'+imlist[i].replace('\n','')+'.jpg'
        print(imlist)
        return imlist, rank_dists