Granulosa cells were cultured in different DSAE concentrations in 24-well plate.
Then 100 cells were counted and the percent of each group were recorded in different DSAE concentration.
In experimental groups DSAE at concentrations of 5, 10, 50, 100, 250, 500, 1000 and 10000 [micro]lg/ml were added to the culture media.
Cell viability assays showed that 10 to 1000 [micro]g/ml of DSAE concentrations did not significantly affect granulosa cells, which remained the same as those in control culture.
Chromatin condensation of 10, 50 and 100 [micro]g/ml DSAE concentrations were nearly the same as control culture.
After treatment with different doses of DSAE, granulosa cells showed a significantly higher percentage of high lipid droplets cells but a lower percentage of the medium and low lipid droplet cells compared to the control culture.
In order to force the hidden layer to learn more robust features and prevent it from simply discovering the sparsity, we train a DSAE to reconstruct the input from a corrupted version of it, which is an extension of SAE.
In this paper, we view DSAE as a "feature extractor" that takes training data X and outputs a function f : [R.sup.C] [right arrow] [R.sup.K] that can map an input vector [x.sub.i] to a new feature vector via the K features, where K is the number of hidden units of DSAE.
ALGORITHM 1: The training procedure of DSAE. (1) Input: (2) Training set X (3) Weight decay parameter [lambda], weight of sparse penalty term [beta], sparse parameter [rho] (4) Procedure: (5) Initialize parameters ([W.sub.1], [b.sub.1]), ([W.sub.2], [b.sub.2]) (6) Get [bar.[x.sup.i]] by stochastic corrupting the input vector [x.sup.i].
In order to extract appropriate and sufficient features from training and testing images, convolution is utilized to construct a locally connected DSAE networks.
(2) Then, the dataset X is fed into a K-hidden-unit network, which is used for unsupervised feature learning of K feature extractors, according to the DSAE model.
(3) After the unsupervised feature learning, convolution is utilized to construct a locally connected DSAE networks.