L-CNN: A Lattice cross-fusion strategy for multistream convolutional neural networks
This addresses the need for better fusion methods in image classification networks, but it is incremental as it builds on existing architectures like AlexNet.
The paper tackled the problem of improving multistream convolutional neural networks by proposing a Lattice Cross Fusion strategy that crosses signals before pooling layers, resulting in a 46% performance improvement over a baseline single stream network on a modified CIFAR-10 dataset.
This paper proposes a fusion strategy for multistream convolutional networks, the Lattice Cross Fusion. This approach crosses signals from convolution layers performing mathematical operation-based fusions right before pooling layers. Results on a purposely worsened CIFAR-10, a popular image classification data set, with a modified AlexNet-LCNN version show that this novel method outperforms by 46% the baseline single stream network, with faster convergence, stability, and robustness.