Image recommendation based on keyword relevance using absorbing Markov chain and image features

Sejal, D.; Rashmi, V.; Venugopal, K R; Iyengar, S. S.; Patnaik, L. M.

doi:10.1007/s13735-016-0104-9

Cited by 16 publications

(8 citation statements)

References 34 publications

(39 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…CDLIR [69], AMCI [75], IRMSR [76] and CRHOG [71] show mAP between 0.66 and 0.76. IKAMC [74] reports lowest mAP as 0.64. Table 5.…”

Section: Results Of the Corel-1000 Dataset With Existing State-of-thementioning

confidence: 93%

“…To test the effectiveness and accuracy of the proposed method, the results of the Corel-1000 dataset were compared with the existing state-of-the-art methods. The existing methods include CDLIR [69], CBSSC [70], CRHOG [71], GRMCB [72], RLMIR [73], IKAMC [74], AMCI [75] and IRMSR [76]. A graphical representation of average precision of the proposed method as compared with existing state-of-the-art methods is shown in Figure 17.…”

Section: Results Of the Corel-1000 Dataset With Existing State-of-thementioning

confidence: 99%

See 1 more Smart Citation

Deep Learning Using Symmetry, FAST Scores, Shape-Based Filtering and Spatial Mapping Integrated with CNN for Large Scale Image Retrieval

et al. 2020

View full text Add to dashboard Cite

This article presents symmetry of sampling, scoring, scaling, filtering and suppression over deep convolutional neural networks in combination with a novel content-based image retrieval scheme to retrieve highly accurate results. For this, fusion of ResNet generated signatures is performed with the innovative image features. In the first step, symmetric sampling is performed on the images from the neighborhood key points. Thereafter, the rotated sampling patterns and pairwise comparisons are performed, which return image smoothing by applying standard deviation. These values of smoothed intensity are calculated as per local gradients. Box filtering adjusts the results of approximation of Gaussian with standard deviation to the lowest scale and suppressed by non-maximal technique. The resulting feature sets are scaled at various levels with parameterized smoothened images. The principal component analysis (PCA) reduced feature vectors are combined with the ResNet generated feature. Spatial color coordinates are integrated with convolutional neural network (CNN) extracted features to comprehensively represent the color channels. The proposed method is experimentally applied on challenging datasets including Cifar-100 (10), Cifar-10 (10), ALOT (250), Corel-10000 (10), Corel-1000 (10) and Fashion (15). The presented method shows remarkable results on texture datasets ALOT with 250 categories and fashion (15). The proposed method reports significant results on Cifar-10 and Cifar-100 benchmarks. Moreover, outstanding results are obtained for the Corel-1000 dataset in comparison with state-of-the-art methods.

show abstract

“…CDLIR [69], AMCI [75], IRMSR [76] and CRHOG [71] show mAP between 0.66 and 0.76. IKAMC [74] reports lowest mAP as 0.64. Table 5.…”

Section: Results Of the Corel-1000 Dataset With Existing State-of-thementioning

confidence: 93%

Section: Results Of the Corel-1000 Dataset With Existing State-of-thementioning

confidence: 99%

Deep Learning Using Symmetry, FAST Scores, Shape-Based Filtering and Spatial Mapping Integrated with CNN for Large Scale Image Retrieval

et al. 2020

View full text Add to dashboard Cite

show abstract

“…(1) Selection and extraction of image features 1) Image color feature extraction Color feature is the most widely used visual feature in image recommendation and retrieval. Compared with other visual features, color is highly correlated with the scene objects contained in the image and has high robustness [18]. Therefore, color is taken as one of the extracted features in the recommendation algorithm of content personalization based on image.…”

Section: Image Recommendation Algorithm Based On Implicit Support mentioning

confidence: 99%

Image Recommendation Algorithm Based on Deep Learning

Yin

Zhang

2020

IEEE Access

View full text Add to dashboard Cite

With the rapid development of network technology, the number of digital images is growing at an alarming rate, people's demand for information gradually from text into images. However, it is very difficult for users to quickly find the images they are interested in from the large number of image libraries. The purpose of this paper is to study the image recommendation algorithm based on deep learning. In this paper, image classification algorithm is firstly studied. LReLU-Softplus activation function is formed by combining LReLU function and Softplus function, and CNN is improved. Then, an image retrieval model based on local sensitive hash algorithm is proposed in this paper. This model calculates the distance in hamming space for the binary hash code generated by mapping. Euclidean distance is calculated inside the result set after similarity measurement to improve the accuracy, and the image retrieval model is constructed. Then, an image recommendation model based on implicit support vector machine (SVM) is proposed in this paper. This image recommendation method combines image text information and image content information. The experimental results show that the proposed image recommendation model can meet the practical needs. In this paper, the overlap rate between the lsvm-based recommendation model and the human recommendation algorithm was tested, and the coincidence degree of the two recommended images reached 88%.

show abstract

“…The methodology uses Cosine Similarity measure for computing the semantic similarity measure. Sejal et al (2016) have proposed a unique strategy for web image recommendation by computing the relevance of keywords incorporating Markov Chain and features of images. Image ranking is mainly based on computing the relevance probability of keywords that are annotated from query input as well as the web log.…”

Section: Related Workmentioning

confidence: 99%

An intelligent inventive system for personalised webpage recommendation based on ontology semantics

Deepak

Ahmed²,

Skanda

2019

IJISTA

View full text Add to dashboard Cite

Owing to the information diversity in the web and its dynamically changing contents, extraction of relevant information from the web is a huge challenge. With the World Wide Web transforming into a more organised semantic web, the incorporation of semantic techniques to retrieve relevant information is highly necessary. In this paper, a dynamic ontology alignment technique for recommending relevant webpages is proposed. The strategy focuses on knowledge tree construction by computing the semantic similarity between the query terms as well as the ontological entities. Furthermore, the semantic similarity is again computed between nodes of the constructed knowledge tree and URLs in the URL repository to recommend relevant webpages. The dynamic ontology alignment by computing their respective semantic similarity constitutes Ontology Semantics. Personalisation is achieved by prioritisation of webpages by content-based analysis of the users' web usage data. An overall accuracy of 87.73% is achieved by the proposed approach.

show abstract

Image recommendation based on keyword relevance using absorbing Markov chain and image features

Cited by 16 publications

References 34 publications

Deep Learning Using Symmetry, FAST Scores, Shape-Based Filtering and Spatial Mapping Integrated with CNN for Large Scale Image Retrieval

Deep Learning Using Symmetry, FAST Scores, Shape-Based Filtering and Spatial Mapping Integrated with CNN for Large Scale Image Retrieval

Image Recommendation Algorithm Based on Deep Learning

An intelligent inventive system for personalised webpage recommendation based on ontology semantics

Contact Info

Product

Resources

About