Self Similarity Wide-Joins for Near-Duplicate Image Detection

Carvalho, Luiz Olmes; Santos, Lucio F. D.; Oliveira, Willian D.; Traina, Agma J. M.; Traina, Caetano

doi:10.1109/ism.2015.114

Cited by 6 publications

(6 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Furthermore, the data duplication technique [46], [47], [48], [49] optimizes database usage by reducing subsequent data match and comparing to bits of help reduce electricity consumption which can increase heat production. Another technique is Similarity Join Operator Technique [42], [50] that evaluate image similarity based on absolute ratios, whereas, Fourier Mellin Transform technique [51] helps image detection via image rotation, image scaling and invariant changes found in the image, followed by the Haar-wavelet technique [52], [53] that assist to extract features vector that is included in the image in order to find the Manhattan distance of the object in the image. Furthermore, kernel hashing technique [54] helps to detect the near-duplicate image by examining the various features contained within the image to detect the differences of each image and convert it to binary of images to place within the kernel space.…”

Section: B Image Features Extractionmentioning

confidence: 99%

Visualising Image Data through Image Retrieval Concept using a Hybrid Technique: Songket Motif’s

Yusof¹,

Ismail²,

Majid³

2019

IJACSA

View full text Add to dashboard Cite

It has been proven that the massive dataset is strictly complex in Content Based image Retrieval (CBIR) because the present strategies in CBIR might have faced difficulties in feature extraction of the images. Moreover, technological constraints encountered in the analysis and extraction of the image arrays are how the system customizes the primitive geometric structures known as polygonal approximations structure. Hence, this study has discovered that image feature extraction is utilized by applying the Principal Component Analysis (PCA) technique, which is primarily based on the matrix of image representation that will enlarge the similarity of detection. The PCA approach needs to be enhanced resulting from the lack of the extraction of features in songket motives images. Therefore, this study proposes a new hybrid model that will integrate PCA with geometric techniques for image feature extraction to increase the recall and precision result. This paper employs the use of a qualitative experimental design model that involves three phases of activities. First, the analysis and design phase, secondly is a development phase, and lastly is the testing and evaluation phase. This paper focuses on those two phases in terms of design and development phases. The outcome process of the empirical phase is followed by designing the algorithm and model based on the result of literature review. This study has found that the hybrid between the principal component analysis model and the geometry technique will help to reduce the problems faced by the basic engineering technique model, which is the constraint in analysing and extracting the image features to customize the geometric primitive structure.

show abstract

Section: B Image Features Extractionmentioning

confidence: 99%

Visualising Image Data through Image Retrieval Concept using a Hybrid Technique: Songket Motif’s

Yusof¹,

Ismail²,

Majid³

2019

IJACSA

View full text Add to dashboard Cite

show abstract

“…• Detecção de quase-duplicatas (near-duplicate detection) [85,18,21]: a detecção de objetos quase duplicados é importante em diversos contextos. Por exemplo, um sistema de arquivamento de informações médicas como o PACS (Picture Archiving and Communication Systems), é capaz de armazenar milhares de imagens de exames.…”

Section: Motivação E Problemáticaunclassified

“…1. Introdução que executa a consulta [62,38,18]. Apesar dos operadores de junção por similaridade realizarem a recuperação dos objetos mais similares, o quinto problema destes operadores é que, em geral, eles retornam um conjunto resposta de cardinalidade elevada.…”

unclassified

Operadores físicos binários para consultas por similaridade em SGBDR

Carvalho¹

Self Cite

View full text Add to dashboard Cite

Ao Senhor Deus, por conceder força e inspiração para a realização deste trabalho. A meus pais, Maria Lades da Silva Carvalho e Ivo Guimarães de Carvalho, por todo o amor, apoio e compreensão durante toda a vida. Ao Prof. Dr. Caetano Traina Jr., por sua orientação sólida, precisa e segura, pelo aprendizado e amadurecimento de diversos conceitos, pela disponibilidade e empenho manifestados. À Profa. Dra. Agma Juci Machado Traina, por seu incentivo, atenção, apoio e encorajamento. Aos amigos para todo o sempre, Christian César Bones e Clélia Junko Kinzú Dimário ("Lia"), por se tornarem parte da minha família e me acolherem na de vocês. Obrigado por estarem sempre presentes ao longo desta caminhada. Sinto falta de vocês por perto! Aos companheiros cujos caminhos se cruzaram ao meu diante de um ideal em comum, e, por todos os momentos compartilhados, a saudade há de ficar. Em especial, Lúcio Fernandes Dutra Santos, Marcos Vinícius Naves Bedo e Willian Dener de Oliveira; sempre em ordem alfabética, como as coisas devem ser. A uma das pessoas mais fantásticas com quem tive o privilégio de conviver, o amigo Enzo Seraphim. Àqueles cujo primor e esmero despontam como fagulhas de ânimo nos mais turvos dias, inspirando e motivando, Fernando Aparecido de Assis e Kleber Mazione Lima Ferreira.

show abstract

“…Caso verdade, o elemento candidato é influenciado por algum elemento em R e é marcado como um elemento influenciado podendo ser podado da resposta final (linhas 5-7). Por outro lado, se candidato não é influenciado por nenhum outro elemento em R div , então candidato é inserido no conjunto de resultado (linhas [11][12]. É importante notar que a quantidade de elementos selecionados depende da distribuição dos dados no espaço de busca ao redor do elemento s i ∈ S. Tal processo se repete até que todo conjunto de entrada R temp tenha sido avaliado (linha 3).…”

unclassified

“…Por outro lado, sempre que um candidato s c não faz parte de outro grupo de elementos muito similares em R, ele é inserido no conjunto de resultado como um representativo (linhas [12][13][14] e a influência é novamente ajustada de acordo com o posicionamento relativo e a quantidade de elementos no resultado, utilizando a heurística CB. Como exemplo, o elemento mais próximo de s q é sempre considerado o primeiro elemento representativo do conjunto e a sua distância ao elemento de busca é utilizada para definir o tamanho do grupo e a área de influência (ξ div1 = δ(s 1 , s q )).…”

unclassified

Similaridade em big data

Santos¹

View full text Add to dashboard Cite

The data being collected and generated nowadays increase not only in volume, but also in complexity, requiring new query operators. Health care centers collecting image exams and remote sensing from satellites and from earth-based stations are examples of application domains where more powerful and flexible operators are required. Storing, retrieving and analyzing data that are huge in volume, structure, complexity and distribution are now being referred to as big data. Representing and querying big data using only the traditional scalar data types are not enough anymore. Similarity queries are the most pursued resources to retrieve complex data, but until recently, they were not available in the Database Management Systems. Now that they are starting to become available, its first uses to develop real systems make it clear that the basic similarity query operators are not enough to meet the requirements of the target applications. The main reason is that similarity is a concept formulated considering only small amounts of data elements. Nowadays, researchers are targeting handling big data mainly using parallel architectures, and only a few studies exist targeting the efficacy of the query answers. This Ph.D. work aims at developing variations for the basic similarity operators to propose better suited similarity operators to handle big data, presenting a holistic vision about the database, increasing the effectiveness of the provided answers, but without causing impact on the efficiency on the searching algorithms. To achieve this goal, four mainly contributions are presented: The first one was a result diversification model that can be applied in any comparison criteria and similarity search operator. The second one focused on defining sampling and grouping techniques with the proposed diversification model aiming at speeding up the analysis task of the result sets. The third contribution concentrated on evaluation methods for measuring the quality of diversified result sets. Finally, the last one defines an approach to integrate the concepts of visual data mining and similarity with diversity searches in content-based retrieval systems, allowing a better understanding of how the diversity property is applied in the query process.

show abstract

Self Similarity Wide-Joins for Near-Duplicate Image Detection

Cited by 6 publications

References 7 publications

Visualising Image Data through Image Retrieval Concept using a Hybrid Technique: Songket Motif’s

Visualising Image Data through Image Retrieval Concept using a Hybrid Technique: Songket Motif’s

Operadores físicos binários para consultas por similaridade em SGBDR

Similaridade em big data

Contact Info

Product

Resources

About