Dynamic Scale Inference by Entropy Minimization

Abstract

Given the variety of the visual world there is not one true scale for recognition: objects may appear at drastically different sizes across the visual field. Rather than enumerate variations across filter channels or pyramid levels, dynamic models locally predict scale and adapt receptive fields accordingly. The degree of variation and diversity of inputs makes this a difficult task. Existing methods either learn a feedforward predictor, which is not itself totally immune to the scale variation it is meant to counter, or select scales by a fixed algorithm, which cannot learn from the given task and data. We extend dynamic scale inference from feedforward prediction to iterative optimization for further adaptivity. We propose a novel entropy minimization objective for inference and optimize over task and structure parameters to tune the model to each input. Optimization during inference improves semantic segmentation accuracy and generalizes better to extreme scale variations that cause feedforward dynamic inference to falter.

References in corpus (8)

• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• On Calibration of Modern Neural Networks
• Unsupervised Domain Adaptation with Residual Transfer Networks
• Some Improvements on Deep Convolutional Neural Network Based Image Classification
• Locally Scale-Invariant Convolutional Neural Networks
• Semi-supervised Domain Adaptation via Minimax Entropy
• Deep Value Networks Learn to Evaluate and Iteratively Refine Structured Outputs
• Blurring the Line Between Structure and Learning to Optimize and Adapt Receptive Fields

Cited by in corpus (4)

• Deformable Kernels: Adapting Effective Receptive Fields for Object Deformation
• Resolution learning in deep convolutional networks using scale-space theory
• Point Cloud Instance Segmentation using Probabilistic Embeddings
• Tailoring: encoding inductive biases by optimizing unsupervised objectives at prediction time