.png){kind=icon}
Return
Content-based image retrieval using local ternary wavelet gradient pattern
.png){kind=icon}
PRE
AI
DOI：10.1007/s11042-019-08039-5.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
With the invention of low cost image capturing devices, image acquisition is no longer a difficult task. The immense popularity of such devices has led to the production of large number of images. For accessing these images easily, efficient indexing and organization of images is required. The field of Content-Based Image Retrieval (CBIR) attempts to achieve this goal. This paper proposes a new multiresolution descriptor- Local Ternary Wavelet Gradient Pattern (LTWGP), for CBIR which combines shape feature and texture feature and utilizes this combination at multiple scales of image to construct feature vector for retrieval. DiscreteWavelet Transform (DWT) coefficients of grayscale image are computed followed by computation of Local Ternary Pattern (LTP) codes of resulting DWT coefficients. Finally, Histogram of Oriented Gradients (HOG) of resulting LTP codes is computed to construct feature vector. The advantage of this technique is that it computes texture through LTP which extracts complex structural arrangement of pixels more efficiently than other texture features such as Local Binary Pattern (LBP), and shape feature through HOG which measures shape of an object as a local feature without performing any segmentation operation. The combination of LTP and HOG is exploited at multiple resolutions of image through DWT to extract varying level of details so that the features left undetected at one level get detected at another level. The combination of LTP, HOG, and DWT constructs efficient feature descriptor which exploits multiple features at more than one resolution of image. The proposed feature descriptor efficiently extracts local directional information obtained through the combination of LTP and HOG at multiple levels of resolution decomposed through DWT. Performance of the proposed method is measured in terms of precision and recall and tested on four benchmark datasets, namely, Corel-1 K, Corel-5 K, Corel-10 K, and GHIM-10 K. The experimental results demonstrate that the proposed method outperforms other state-of-the-art CBIR techniques in terms of precision and recall.
Keywords:
Content-based image retrieval
Discrete wavelet transform
Local ternary pattern
Histogram of oriented gradients
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3
Papers:
2.0W
Citations:
3.2W
Organization
Cited Papers
ShareSaveShareSaveShareSave
ShareSaveContent-based image retrieval using computational visual attention model.png)
PATTERN RECOGNITION
PATTERN RECOGNITIONIF7.6
PREAI
ShareSave
View more