Inception Convolution with Efficient Dilation Search
arXiv:2012.13587
Abstract
As a variant of standard convolution, a dilated convolution can control effective receptive fields and handle large scale variance of objects without introducing additional computational costs. To fully explore the potential of dilated convolution, we proposed a new type of dilated convolution (referred to as inception convolution), where the convolution operations have independent dilation patterns among different axes, channels and layers. To develop a practical method for learning complex inception convolution based on the data, a simple but effective search algorithm, referred to as efficient dilation optimization (EDO), is developed. Based on statistical optimization, the EDO method operates in a low-cost manner and is extremely fast when it is applied on large scale datasets. Empirical results validate that our method achieves consistent performance gains for image recognition, object detection, instance segmentation, human detection, and human pose estimation. For instance, by simply replacing the 3x3 standard convolution in the ResNet-50 backbone with inception convolution, we significantly improve the AP of Faster R-CNN from 36.4% to 39.2% on MS COCO.
References in corpus (11)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Rethinking Atrous Convolution for Semantic Image Segmentation
- Neural Architecture Search with Reinforcement Learning
- R-FCN: Object Detection via Region-based Fully Convolutional Networks
- Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
- Associative Embedding: End-to-End Learning for Joint Detection and Grouping
- CrowdHuman: A Benchmark for Detecting Human in a Crowd
- Understanding and Robustifying Differentiable Architecture Search
- DetNet: A Backbone network for Object Detection
- Computation Reallocation for Object Detection
- Powering One-shot Topological NAS with Stabilized Share-parameter Proxy