Autonomous deep learning: A genetic DCNN designer for image classification

•Search deep convolutional neural networks automatically for image classification.•Encode a deep convolutional neural network’s architecture into an integer string.•Evolve a population of DCNN architectures using genetic evolutionary operations.•Obtained DCNNs achieved satisfying results with less layers than pre-trained models.•Image classification using neither handcrafted features nor handcrafted networks.