Multimodal Re-ranking of Product Image Search Results

Santos, Joyce Miranda dos; Cavalcanti, João M. B.; Saraiva, Patricia Correia; Moura, Edleno Silva de

doi:10.1007/978-3-642-36973-5_6

Cited by 6 publications

(14 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [11], a visual reranking method was introduced for web image search with the help of intention categorization. Currently, an up-to-date work of [13] introduced a multimodal reranking method relying on category and textual information with a promising result. Our idea in this paper is similar to this work in some sense.…”

Section: Query Expansion and Rerankingmentioning

confidence: 97%

“…Then, move the results in top k answers that meet the category requirement to the front. However, there are two drawbacks in [13]: (1) Lack sufficient study of category prediction; (2) The category information for each dataset image is required, which is unavailable in unsupervised situations.…”

Section: Query Expansion and Rerankingmentioning

confidence: 99%

“…Although achieving a perfect retrieval result is challenging for both image and 3D retrieval, another distinctive technique for improving the query quality is more feasible, which has been verified in many applications [13,16]. By utilizing query expansion or reranking method, it is possible to improve the retrieval performance.…”

Section: Query Expansion and Rerankingmentioning

confidence: 99%

“…The shortlist of a retrieval result contains interesting informations and lots of retrieval works [10,13] use them for reranking. The shortlist here can be the result of any retrieval method, such as KNN(k-nearest neighbor) which, without loss of generality, is used in this section.…”

Section: Retrieval Guidance (Rg)mentioning

confidence: 99%

“…In usual, the top queries or some learned informations are exploited for reranking. Specifically, [13] used the category information to improve the query performance and [16] adopted the spatial verification method to exclude incorrect results, and both of them have lifted the retrieval quality. However, these methods did not explore the usage of potential categories results further, which has limited their performance.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Exploration in improving retrieval quality and robustness for deformable non-rigid 3D shapes

Kuang¹,

Jiang

et al. 2014

Multimed Tools Appl

View full text Add to dashboard Cite

Improving query quality and robustness is a hot topic in information and image retrieval field, which has resulted in many interesting works. To address the same problem for deformable non-rigid 3D shape retrieval, two topics are considered in this paper. The first one we discussed is shape representation, which is related to feature extraction and fusion. For feature extraction, we create a global feature to achieve a coarser-scale shape appearance description. Then, to alleviate the drawbacks of retrieval by single feature, we develop a novel fusion method for multiple feature fusion, which turns out to be superior to weighted sum approach with a low complexity. The second topic studied in this paper is to further refine the retrieval results by introducing a new retrieval guidance algorithm based on category prediction. To evaluate the proposed methods, experiments on three popular nonrigid datasets are carried out. The evaluation results suggest that our shape representation method has achieved state-of-the-art performance. Then, by adjusting the retrieval results of existing methods, our retrieval guidance algorithm has promoted the accuracy with nice effects.

show abstract

Section: Query Expansion and Rerankingmentioning

confidence: 97%

Section: Query Expansion and Rerankingmentioning

confidence: 99%

Section: Query Expansion and Rerankingmentioning

confidence: 99%

Section: Retrieval Guidance (Rg)mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations