Compact Global Descriptor for Neural Networks
This work addresses the problem of high computational cost in long-range dependency modeling for computer vision tasks, offering a more efficient solution for researchers and practitioners in the field, though it appears incremental as it builds on existing mechanisms.
The paper tackles the computational inefficiency of existing long-range dependency modeling methods in CNNs by introducing a lightweight global descriptor that provides global features to subsequent convolutions with minimal computational cost and parameters, achieving state-of-the-art performance with significant reductions in extra computing cost.
Long-range dependencies modeling, widely used in capturing spatiotemporal correlation, has shown to be effective in CNN dominated computer vision tasks. Yet neither stacks of convolutional operations to enlarge receptive fields nor recent nonlocal modules is computationally efficient. In this paper, we present a generic family of lightweight global descriptors for modeling the interactions between positions across different dimensions (e.g., channels, frames). This descriptor enables subsequent convolutions to access the informative global features with negligible computational complexity and parameters. Benchmark experiments show that the proposed method can complete state-of-the-art long-range mechanisms with a significant reduction in extra computing cost. Code available at https://github.com/HolmesShuan/Compact-Global-Descriptor.