NAS-FCOS: Efficient Search for Object Detection Architectures

TL;DR

This paper introduces NAS-FCOS, an efficient neural architecture search method for object detection that finds high-performing architectures within 4 days, surpassing existing models in accuracy with similar complexity.

Contribution

The paper presents a tailored reinforcement learning approach to efficiently search for feature pyramid and prediction head architectures for FCOS, achieving state-of-the-art results.

Findings

Surpassed state-of-the-art detection models by 1.0% to 5.4% AP on COCO

Found architectures within 4 days using 8 V100 GPUs

Achieved high accuracy with comparable computational complexity

Abstract

Neural Architecture Search (NAS) has shown great potential in effectively reducing manual effort in network design by automatically discovering optimal architectures. What is noteworthy is that as of now, object detection is less touched by NAS algorithms despite its significant importance in computer vision. To the best of our knowledge, most of the recent NAS studies on object detection tasks fail to satisfactorily strike a balance between performance and efficiency of the resulting models, let alone the excessive amount of computational resources cost by those algorithms. Here we propose an efficient method to obtain better object detectors by searching for the feature pyramid network (FPN) as well as the prediction head of a simple anchor-free object detector, namely, FCOS [36], using a tailored reinforcement learning paradigm. With carefully designed search space, search algorithms, and strategies for evaluating network quality, we are able to find top-performing detection architectures within 4 days using 8 V100 GPUs. The discovered architectures surpass state-of-the-art object detection models (such as Faster R-CNN, Retina-Net and, FCOS) by 1.0% to 5.4% points in AP on the COCO dataset, with comparable computation complexity and memory footprint, demonstrating the efficacy of the proposed NAS method for object detection. Code is available at https://github.com/Lausannen/NAS-FCOS.