OPANAS: One-Shot Path Aggregation Network Architecture Search for Object Detection
Tingting Liang, Yongtao Wang, Zhi Tang, Guosheng Hu, Haibin Ling

TL;DR
OPANAS introduces an efficient neural architecture search method for feature pyramid networks, significantly enhancing object detection accuracy and speed while reducing search costs, outperforming existing NAS-based FPN designs.
Contribution
The paper proposes a novel one-shot NAS approach with a new search space and an efficient evolutionary search method for FPNs, improving detection performance and search efficiency.
Findings
OPANAS reduces search cost to 4 GPU days on MS-COCO.
It improves detection accuracy by 2.3-3.2% mAP over baseline FPNs.
Achieves a new state-of-the-art trade-off with 52.2% mAP at 7.6 FPS.
Abstract
Recently, neural architecture search (NAS) has been exploited to design feature pyramid networks (FPNs) and achieved promising results for visual object detection. Encouraged by the success, we propose a novel One-Shot Path Aggregation Network Architecture Search (OPANAS) algorithm, which significantly improves both searching efficiency and detection accuracy. Specifically, we first introduce six heterogeneous information paths to build our search space, namely top-down, bottom-up, fusing-splitting, scale-equalizing, skip-connect and none. Second, we propose a novel search space of FPNs, in which each FPN candidate is represented by a densely-connected directed acyclic graph (each node is a feature pyramid and each edge is one of the six heterogeneous information paths). Third, we propose an efficient one-shot search method to find the optimal path aggregation architecture, that is, we…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Neural Network Applications · Visual Attention and Saliency Detection · Video Surveillance and Tracking Methods
MethodsRegion Proposal Network · Convolution · RoIPool · Average Pooling · Global Average Pooling · Softmax · Faster R-CNN · Batch Normalization · *Communicated@Fast*How Do I Communicate to Expedia? · NAS-FPN
