Abstract | ||
---|---|---|
Though there are some deep learning approaches that have acquired satisfied results on image super-resolution (SR), many lacks intelligence in the real sense. This paper explores and finds a specific correlation between semantic category information and SR procedure, to construct a complete SR reconstruction system which can realize online working and learning. The proposed system flow takes the memory mechanism of human as a reference and is divided into three parts, sensory memory, short-term memory (STM) and longterm memory (LTM). The test evaluation of the experiments shows that this framework can be adpoted to obtain a better performance than other traditional and deep learning methods. From these results, it can be concluded that the conventional wisdom of divide and conquer is also applicable to deep learning based methods. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1109/ROBIO.2018.8665116 | 2018 IEEE International Conference on Robotics and Biomimetics (ROBIO) |
Keywords | Field | DocType |
Semantics,Deep learning,Image reconstruction,Training,Brain modeling,Spatial resolution | Iterative reconstruction,Test evaluation,Control engineering,Artificial intelligence,Divide and conquer algorithms,Deep learning,Engineering,Sensory memory,Superresolution,Image resolution,Semantics | Conference |
ISBN | Citations | PageRank |
978-1-7281-0377-8 | 0 | 0.34 |
References | Authors | |
0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ziyang Liu | 1 | 840 | 42.54 |
Weihai Chen | 2 | 4 | 3.45 |
Xingming Wu | 3 | 43 | 13.16 |
Haosong Yue | 4 | 2 | 3.10 |
Z. Li | 5 | 1578 | 164.19 |