Object Retrieval in Images using SIFT and R-CNN

Amitha, I. C.; Narayanan, N. K.

doi:10.1109/icitiit49094.2020.9071557

Cited by 9 publications

(5 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A combination of techniques that can handle texture and color features is needed to improve accuracy for image types with complex textures. Color, texture, shape, spatial information Combining features improves retrieval accuracy [7], [8], [10], [11], [13], [15], [19] Small feature dimensions [8], [15], [16], [21] Faster computation time [8], [9], [12], [15] Can handle background interference and scene rotation [14], [53] Not suitable for complex fabric images [7], [9], [11] Use of a single feature has low retrieval accuracy [6], [8], [11], [12], [17], [18], [28]. Retrieval accuracy is lower than with CNNs…”

Section: Comparison Of Traditional and Cnn Feature Extrationmentioning

confidence: 99%

Content-based image retrieval for fabric images: A survey

Tena¹,

Hartanto²,

Ardiyanto³

2023

Preprint

View full text Add to dashboard Cite

In recent years, a great deal of research has been conducted in the area of fabric image retrieval, especially the identification and classification of visual features. One of the challenges associated with the domain of content-based image retrieval (CBIR) is the semantic gap between low-level visual features and high-level human perceptions. Generally, CBIR includes two main components, namely feature extraction and similarity measurement. Therefore, this research aims to determine the content-based image retrieval for fabric using feature extraction techniques grouped into traditional methods and convolutional neural networks (CNN). Traditional descriptors deal with low-level features, while CNN addresses the high-level, called semantic features. Traditional descriptors have the advantage of shorter computation time and reduced system requirements. Meanwhile, CNN descriptors, which handle high-level features tailored to human perceptions, deal with large amounts of data and require a great deal of computation time. In general, thefeatures of a CNN's fully connected layers are used for matching query and database images. In several studies, the extracted features of the CNN's convolutional layer were used for image retrieval. At the end of the CNN layer, hash codes are added to reduce search time.

show abstract

Section: Comparison Of Traditional and Cnn Feature Extrationmentioning

confidence: 99%

Content-based image retrieval for fabric images: A survey

Tena¹,

Hartanto²,

Ardiyanto³

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…(4) ImageDec: After the users obtain the retrieval results, the users can decrypt the secret image with their own private key pk u to decrypt C K = (c k1 , c k2 ) and obtain the retrieval results M k as shown in Equation (18).…”

Section: Decryptmentioning

confidence: 99%

“…This characteristic adds complexity to the retrieval process. While methods like R-CNN can be employed to detect object boundaries, define regions of interest (RoIs), and enhance convolutional neural network features for individual regions to facilitate classification, as explored by Amitha et al [18], there is a shortage of research on retrieving singleobject images (small images) within encrypted environments. Moreover, no existing research tackles the challenge of retrieving multi-object images (large images) from single-object images (small images) while under encryption.…”

Section: Introductionmentioning

confidence: 99%

Fine-Grained Encrypted Image Retrieval in Cloud Environment

Chen,

Huang

2023

Mathematics

View full text Add to dashboard Cite

With the growing emphasis on privacy awareness, there is an increasing demand for privacy-preserving encrypted image retrieval and secure image storage on cloud servers. Nonetheless, existing solutions exhibit certain shortcomings regarding retrieval accuracy, the capacity to search large images from smaller ones, and the implementation of fine-grained access control. Consequently, to rectify these issues, the YOLOv5 technique is employed for object detection within the image, capturing them as localized images. A trained convolutional neural network (CNN) model extracts the feature vectors from the localized images. To safeguard the encrypted image rules from easy accessibility by third parties, the image is encrypted using ElGamal. In contrast, the feature vectors are encrypted using the skNN method to achieve ciphertext retrieval and then upload this to the cloud. In pursuit of fine-grained access control, a role-based multinomial access control technique is implemented to bestow access rights to local graphs, thereby achieving more nuanced permission management and heightened security. The proposed scheme introduces a comprehensive cryptographic image retrieval and secure access solution, encompassing fine-grained access control techniques to bolster security. Ultimately, the experiments are conducted to validate the proposed solution’s feasibility, security, and accuracy. The solution’s performance across various facets is evaluated through these experiments.

show abstract

“…A combination of techniques that can handle texture and color features is needed to improve accuracy for image types with complex textures. [7], [8], [10], [11], [13], [15], [19] Small feature dimensions [8], [15], [16], [21] Faster computation time [8], [9], [12], [15] Can handle background interference and scene rotation [14], [53] Not suitable for complex fabric images [7], [9], [11] Use of a single feature has low retrieval accuracy [6], [8], [11], [12], [17], [18], [28]. Retrieval accuracy is lower than with CNNs…”

Section: Comparison Of Traditional and Cnn Feature Extrationmentioning

confidence: 99%

Content-based image retrieval for fabric images: A survey

Tena

Hartanto

Ardiyanto

2021

IJEECS

View full text Add to dashboard Cite

In <span>recent years, a great deal of research has been conducted in the area of fabric image retrieval, especially the identification and classification of visual features. One of the challenges associated with the domain of content-based image retrieval (CBIR) is the semantic gap between low-level visual features and high-level human perceptions. Generally, CBIR includes two main components, namely feature extraction and similarity measurement. Therefore, this research aims to determine the content-based image retrieval for fabric using feature extraction techniques grouped into traditional methods and convolutional neural networks (CNN). Traditional descriptors deal with low-level features, while CNN addresses the high-level, called semantic features. Traditional descriptors have the advantage of shorter computation time and reduced system requirements. Meanwhile, CNN descriptors, which handle high-level features tailored to human perceptions, deal with large amounts of data and require a great deal of computation time. In general, the features of a CNN's fully connected layers are used for matching query and database images. In several studies, the extracted features of the CNN's convolutional layer were used for image retrieval. At the end of the CNN layer, hash codes are added to reduce </span>search time.

show abstract

Object Retrieval in Images using SIFT and R-CNN

Cited by 9 publications

References 11 publications

Content-based image retrieval for fabric images: A survey

Content-based image retrieval for fabric images: A survey

Fine-Grained Encrypted Image Retrieval in Cloud Environment

Content-based image retrieval for fabric images: A survey

Contact Info

Product

Resources

About