In this study, we evaluated four binarization methods. Locality-Sensitive Hashing (LSH), Iterative Quantization (ITQ), Kernel-based Supervised Hashing (KSH), and Supervised Discrete Hashing (SDH) on the ODIR dataset using deep feature embeddings. Experimental results show that SDH achieved the best performance, with an mAP@100 of 0.9184 using only 32-bit codes, outperforming LSH, ITQ, and KSH. Compared with prior studies, our method proved highly competitive: Fang et al. reported 0.7528 (Fundus-iSee, 48 bits) and 0.8856 (ASOCT-Cataract, 48 bits), while Wijesinghe et al. achieved 94.01 (KVASIR, 256 bits). Despite using significantly fewer bits, our SDH-based framework reached retrieval accuracy close to the state-of-the-art. These findings demonstrate that SDH is the most effective approach among those tested, offering a practical balance of accuracy, storage, and efficiency for medical image retrieval and device inventory management.

翻译：本研究在ODIR数据集上，使用深度特征嵌入评估了四种二值化方法：局部敏感哈希（LSH）、迭代量化（ITQ）、基于核的监督哈希（KSH）和监督离散哈希（SDH）。实验结果表明，SDH取得了最佳性能，仅使用32位编码即获得0.9184的mAP@100，优于LSH、ITQ和KSH。与先前研究相比，我们的方法展现出高度竞争力：Fang等人报告了0.7528（Fundus-iSee，48位）和0.8856（ASOCT-Cataract，48位），而Wijesinghe等人实现了94.01（KVASIR，256位）。尽管使用的比特数显著更少，我们基于SDH的框架达到了接近最先进水平的检索精度。这些发现证明，在所测试的方法中，SDH是最有效的途径，为医学图像检索和设备库存管理提供了精度、存储和效率之间的实用平衡。