Image Retrieval Algorithm based on Compressed Sensing and Dictionary Learning Methods
Xu Yang, Ruiting Sun, Chenning Yu
Available Online April 2017.
- https://doi.org/10.2991/icmmct-17.2017.123How to use a DOI?
- Image retrieval, semantic information, Nyquist sampling rate, weighted distance, precise image retrieving, cognitive theory, bi-cubic interpolation (BI).
- Image retrieval is based on the description of image content. The content of the image can be divided into two categories: visual content and information content. The visual content corresponds to the physical representation of the image, such as color, shape, texture. The information content corresponds to the semantic representation of the image, such as the theme, characters, and scenes. This paper discusses the image retrieval algorithm based on compressed sensing and dictionary learning methods. Compressed sensing, also known as compression sampling or compressed sensing, is a new sampling theory which acquires the correct signal a sampling speed far less than the Nyquest sampling rate. Compressed sensing technology randomly samples the signal through the development of signal sparse features, and then reconstructs the signal through nonlinear perfect signal reconstruction algorithms. In this paper, compressed sensing theory is applied to image retrieval in the process of feature extraction and matching. Combining with the optimized dictionary learning, a new retrieval model reconstruction algorithm is established. Experiments show that our algorithm can achieve high compression ratio through compression perception of linear measurement process. Using weighted distance method to calculate the similarity of measured value of the image features, the precise image retrieving is realized.
- Open Access
- This is an open access article distributed under the CC BY-NC license.
Cite this article
TY - CONF AU - Xu Yang AU - Ruiting Sun AU - Chenning Yu PY - 2017/04 DA - 2017/04 TI - Image Retrieval Algorithm based on Compressed Sensing and Dictionary Learning Methods BT - Proceedings of the 2017 5th International Conference on Machinery, Materials and Computing Technology (ICMMCT 2017) PB - Atlantis Press SP - 587 EP - 594 SN - 2352-5401 UR - https://doi.org/10.2991/icmmct-17.2017.123 DO - https://doi.org/10.2991/icmmct-17.2017.123 ID - Yang2017/04 ER -