Dual Skipping Networks
This work addresses object categorization challenges in computer vision, but it appears incremental as it builds on existing neuroscience-inspired architectures without claiming major breakthroughs.
The paper tackles coarse-to-fine object categorization by introducing a dual skipping network with two branches for handling coarse and fine-grained tasks, using a layer-skipping mechanism to predict which layers to skip during testing, achieving promising results on benchmarks.
Inspired by the recent neuroscience studies on the left-right asymmetry of the human brain in processing low and high spatial frequency information, this paper introduces a dual skipping network which carries out coarse-to-fine object categorization. Such a network has two branches to simultaneously deal with both coarse and fine-grained classification tasks. Specifically, we propose a layer-skipping mechanism that learns a gating network to predict which layers to skip in the testing stage. This layer-skipping mechanism endows the network with good flexibility and capability in practice. Evaluations are conducted on several widely used coarse-to-fine object categorization benchmarks, and promising results are achieved by our proposed network model.