Image quality assessment using similar scenes as reference
A system for image quality assessment of non-aligned images includes a first deep path portion of a convolutional neural network having a set of parameters and a second deep path portion of the convolutional neural network sharing a set of parameters with the first deep path convolutional neural network. Weights are shared between the first and second deep path convolutional neural networks to support extraction of a same set of features in each neural network pathway. Non-aligned reference and distorted images are respectively provided to the first and second deep paths of the convolutional neural network for processing. A concatenation layer is connected to both the first and second deep paths convolutional neural network, and a fully connected layer is connected to the concatenation layer to receive input from both the first and second deep paths of the convolutional neural network, generating an image quality assessment as a linear regressor and outputting an image quality score.
1. An apparatus comprising:
at least one camera system providing distorted images; and
a computer system including a first deep path convolutional neural network, a second deep path convolutional neural network and a fully connected layer, the first deep path convolutional neural network accessing reference images while using a set of parameters, the second deep path convolutional neural network receiving the distorted images from the at least one camera system, the second deep path convolutional neural network sharing the set of parameters with the first deep path convolutional neural network, the first deep path convolutional neural network and the second deep path convolutional neural network sharing weights to support extraction of a same set of features, and the fully connected layer configured to receive input from both the first deep path convolutional neural network and the second deep path convolutional neural network, generate an image quality assessment as a linear regressor, and output an image quality score.
2. The apparatus of claim 1 , wherein both the first and second deep paths of the convolutional neural network have multiple respective convolution layers, and at least one respective pooling layer.
3. The apparatus of claim 2 , wherein a pooling layer processes a sub-image derived from a reference image or a distorted image.
4. The apparatus of claim 3 , wherein the dimensions of the sub-image are 224 pixels×224 pixels.
5. The apparatus of claim 1 , wherein both the first and second deep paths of the convolutional neural network feed into a concatenation layer to concatenate features from the first and second deep paths of the convolutional neural networks.
6. The apparatus of claim 1 , wherein both the first and second deep paths of the convolutional neural network feed into a concatenation layer to concatenate features from the first and second deep paths of the convolutional neural network and discriminative features are mapped in fully connected layers to generate the image quality assessment as a linear regressor.
7. The apparatus of claim 1 , wherein the image quality score is output by minimizing a Euclidean loss.
8. The apparatus of claim 1 , wherein numeric stability is obtained by local contrast normalization.
9. The apparatus of claim 8 , wherein the local contrast normalization is a data whitening process.
10. The apparatus of claim 1 , wherein the first and second deep paths of the convolutional neural network are trained using a stochastic gradient descent with a standard back-propagation.