On the Selective and Invariant Representation of DCNN for High-Resolution Remote Sensing Image Recognition
This work provides insights into DCNN mechanisms for remote sensing image analysis, which could aid in designing better models for applications in environmental monitoring or urban planning, but it is incremental as it applies existing methods to a specific domain.
The authors investigated why deep convolutional neural networks (DCNNs) achieve high accuracy in recognizing high-resolution remote sensing images by analyzing selective and invariant representations at the neuron level using AlexNet on a large-scale benchmark, finding that these representations are crucial for tasks like classification and detection.
Human vision possesses strong invariance in image recognition. The cognitive capability of deep convolutional neural network (DCNN) is close to the human visual level because of hierarchical coding directly from raw image. Owing to its superiority in feature representation, DCNN has exhibited remarkable performance in scene recognition of high-resolution remote sensing (HRRS) images and classification of hyper-spectral remote sensing images. In-depth investigation is still essential for understanding why DCNN can accurately identify diverse ground objects via its effective feature representation. Thus, we train the deep neural network called AlexNet on our large scale remote sensing image recognition benchmark. At the neuron level in each convolution layer, we analyze the general properties of DCNN in HRRS image recognition by use of a framework of visual stimulation-characteristic response combined with feature coding-classification decoding. Specifically, we use histogram statistics, representational dissimilarity matrix, and class activation mapping to observe the selective and invariance representations of DCNN in HRRS image recognition. We argue that selective and invariance representations play important roles in remote sensing images tasks, such as classification, detection, and segment. Also selective and invariance representations are significant to design new DCNN liked models for analyzing and understanding remote sensing images.