Considerations in Evaluation of Deep Hashing Networks for Information Retrieval System

Deep hashing is a method utilized in information retrieval systems, involving learning hash functions using deep neural networks. Mean Average Precision (mAP), a popular metric for evaluating hashing models, faces critical challenges in providing reliable performance scores. Despite the development of recent metrics like Mean Local Group Average Precision (mL-GAP) and Radius Aware Mean Average Precision (RAMAP), only a limited number of papers evaluated their hashing algorithms using these metrics. In this paper, we compare the performance of common deep hashing models using various evaluation metrics for precise comparison.