An Efficient Approximate kNN Graph Method for Diffusion on Image Retrieval

Abstract

The application of the diffusion in many computer vision and artificial intelligence projects has been shown to give excellent improvements in performance. One of the main bottlenecks of this technique is the quadratic growth of the kNN graph size due to the high-quantity of new connections between nodes in the graph, resulting in long computation times. Several strategies have been proposed to address this, but none are effective and efficient. Our novel technique, based on LSH projections, obtains the same performance as the exact kNN graph after diffusion, but in less time (approximately 18 times faster on a dataset of a hundred thousand images). The proposed method was validated and compared with other state-of-the-art on several public image datasets, including Oxford5k, Paris6k, and Oxford105k.

Genetic Algorithms for the Optimization of Diffusion Parameters in Content-Based Image Retrieval

“…Time mAP genetic algorithms 18787 s 97.32% manual configuration [27] 4 s 97.01% Table 7: Comparison of different approaches for the optimization of the diffusion parameters on Paris6k. Table 7 reports the results of different optimization methods applied on Paris6k.…”

“…Time mAP genetic algorithms 63911 s 94.20% manual configuration [27] 13 s 92.50% Table 8: Comparison of different approaches for the optimization of the diffusion parameters on Oxford105k.…”

“…We follow the idea of Magliani et al [27] that exploits the LSH (Locality Sensitive Hashing) [20] to approximately split the elements in several buckets using the hash table mechanism. This method can reduce the time required for the creation of the kNN graph, maintaining or, in some cases, improving the final retrieval performance obtained after the diffusion application.…”

“…Although the diffusion improves retrieval results, it requires a long time to create the kNN graph necessary for the diffusion application. To solve this issue we follow the technique proposed by Magliani et al [27], that proposes a method for effective and efficient creation of an approximate kNN graph suitable for the application of the diffusion approach. On this graph it is possible to obtain the same or better retrieval performance after diffusion than using a brute-force approach, requiring a shorter computation time.…”

Diffusion Parameters Analysis in a Content-Based Image Retrieval Task for Mobile Vision

“…To do so, it is necessary to create the k-Nearest Neighbors (kNN) graph of the database embeddings, but this task is computationally very demanding. To solve this issue, in [ 14 ], we proposed a technique based on a pipeline for efficient and effective diffusion-based retrieval which relies on an approximate kNN graph. Briefly, the graph is constructed following a divide-and-conquer method, based on unsupervised hashing functions.…”