Adaptive Measurement Network for CS Image Reconstruction
This work addresses the problem of inefficient measurements in CS image reconstruction for applications requiring fast and high-quality recovery, though it is incremental as it builds on existing neural network methods.
The paper tackles the slow speed of conventional compressive sensing (CS) image reconstruction by proposing an adaptive measurement network that learns measurements, resulting in better reconstruction quality compared to random Gaussian measurements under the same measurement rate.
Conventional compressive sensing (CS) reconstruction is very slow for its characteristic of solving an optimization problem. Convolu- tional neural network can realize fast processing while achieving compa- rable results. While CS image recovery with high quality not only de- pends on good reconstruction algorithms, but also good measurements. In this paper, we propose an adaptive measurement network in which measurement is obtained by learning. The new network consists of a fully-connected layer and ReconNet. The fully-connected layer which has low-dimension output acts as measurement. We train the fully-connected layer and ReconNet simultaneously and obtain adaptive measurement. Because the adaptive measurement fits dataset better, in contrast with random Gaussian measurement matrix, under the same measuremen- t rate, it can extract the information of scene more efficiently and get better reconstruction results. Experiments show that the new network outperforms the original one.