Visual Image Search: Feature Signatures or/and Global Descriptors

Lokoč, Jakub; Novák, David; Batko, Michal; Skopal, Tomáš

doi:10.1007/978-3-642-32153-5_13

Cited by 11 publications

(7 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Recall measurement can be used in classification the image I [30]. (6) where is the number of correct images classification, is total number of images of class .…”

Section: Cbir Evaluationmentioning

confidence: 99%

See 1 more Smart Citation

Content-Based Image Retrieval: Survey

Al-Jubouri¹

2019

JEASD

View full text Add to dashboard Cite

Extensive use of digital photographic devices has resulted in large volumes of digital images being acquired and stored in databases. Whether it is for scientific research, medical or social networking, there is a growing demand for effective retrieval of digital images based on their visual content (e.g. colour and texture). Content-Based Image Retrieval systems are developed to meet this demand. However, searching for similar and relevant images from large-scale databases still poses a challenge for Content-Based Image Retrieval systems due to the gap between high-level meaning and low-level visual features. This paper reviews different Content-Based Image Retrieval approaches such as Clustering, Region-of-Interest, Bag-of-Visual-Words, Relevance Feedback, Browsing, and indexing that have been developed to reduce such "Semantic gap" issue. So, the interested researchers can interest to determine which method is benefit to his work.

show abstract

“…Recall measurement can be used in classification the image I [30]. (6) where is the number of correct images classification, is total number of images of class .…”

Section: Cbir Evaluationmentioning

confidence: 99%

“…The average accuracy of classification was 88%. The method in [6] used the kmeans algorithm to cluster 7D feature vectors (SQFD): colour from L * , a * , and b * components, coordinates x and y, contrast X, and entropy value. As a result, an image feature composed of centroids .…”

Section: Clustering Approachmentioning

confidence: 99%

Content-Based Image Retrieval: Survey

Al-Jubouri¹

2019

JEASD

View full text Add to dashboard Cite

show abstract

“…In addition, different approaches are developed to reduce or narrow the semantic gap between high level conceptual meaning and the low-level content-based features. Clustering [9], Region of Interest (ROI) [10], Relevance Feedback (RF) [11], Bag of Visual Words (BOVW) [12] and Browsing [13] are the main existing approaches each of which is an active research area by its own right. All the approaches involve using feature extraction and similarity measures to retrieve the most similar images in a ranked list.…”

Section: Introductionmentioning

confidence: 99%

Untitled

2018

IJEEE

View full text Add to dashboard Cite

Content-Based Image Retrieval (CBIR) is an automatic process of retrieving images that are the most similar to a query image based on their visual content such as colour and texture features. However, CBIR faces the technical challenge known as the semantic gap between high level conceptual meaning and the low-level image based features. This paper presents a new method that addresses the semantic gap issue by exploiting cluster shapes. The method first extracts local colours and textures using Discrete Cosine Transform (DCT) coefficients. The Expectation-Maximization Gaussian Mixture Model (EM/GMM) clustering algorithm is then applied to the local feature vectors to obtain clusters of various shapes. To compare dissimilarity between two images, the method uses a dissimilarity measure based on the principle of Kullback-Leibler divergence to compare pair-wise dissimilarity of cluster shapes. The paper further investigates two respective scenarios when the number of clusters is fixed and adaptively determined according to cluster quality. Experiments are conducted on publicly available WANG and Caltech6 databases. The results demonstrate that the proposed retrieval mechanism based on cluster shapes increases the image discrimination, and when the number of clusters is fixed to a large number, the precision of image retrieval is better than that when the relatively small number of clusters is adaptively determined.

show abstract

“…Visando o combate ao gap semântico em consultas por similaridade, a utilização de combinação de múltiplos descritores tem sido bastante pesquisada e tem demonstrado bons resultados: [6], [7], [8], [9], [10], [11]. Características intrín-secas diferentes tendem a se complementar na representação do dado, bem como é realizada pela percepção humana.Por este motivo a utilização de múltiplos descritores tende a melhorar a capacidade de discriminação dos dados.…”

Section: Introductionunclassified