硬件感知的高效特征融合网络搜索  

Hardware-Aware and Efficient Feature Fusion Network Search

作　　者：郭家明 张蕊 支天 何得园 黄迪 常明 张曦珊 郭崎[1] GUO Jia-Ming;ZHANG Rui;ZHI Tian;HE De-Yuan;HUANG Di;CHANG Ming;ZHANG Xi-Shan;GUO Qi(State Key Laboratory of Computer Architecture,Institute of Computing Technology,Chinese Academy of Sciences,Beijing 100190;Cambricon Technologies,Beijing 100191;University of Chinese Academy of Sciences,Beijing 100049)

机构地区：[1]中国科学院计算技术研究所计算机体系结构国家重点实验室,北京100190 [2]寒武纪科技,北京100191 [3]中国科学院大学,北京100049

出　　处：《计算机学报》2022年第11期2420-2432,共13页Chinese Journal of Computers

基　　金：国家重点研发计划(2017YFA0700900);国家自然科学基金(61925208,62102399,61906179,61732020,U20A20227);中国科学院战略性先导科技专项(XDB32050200);中国科学院稳定支持基础研究领域青年团队计划(YSBR-029);科学探索奖资助.

摘　　要：特征融合网络通过融合多尺度特征来提高目标检测精度,是深度学习目标检测框架中的关键部分.已有的研究工作通过优化融合网络的拓扑结构来提高结果精度,忽略了所需的硬件资源开销以及特征选择和特征融合操作对结果的影响.本文提出了支持多尺度特征融合的注意力感知融合网络(Attention-aware Fusion Network,AFN),通过软硬件协同可实现硬件开销(参数存储、计算时间等)敏感的神经网络自动搜索,从融合网络的特征、路径和操作三方面实现一体化的优化部署.实验结果表明,当主干网络为ResNet50时,在实现相似检测精度时,相比现有最先进的搜索网络NAS-FPN,本文方法的参数量和计算量分别减少29.6%和22.3%,相比现有人工设计网络FPN,本文方法的AP可以提高2.1%.当主干网络为VGG时,相比现有最先进的搜索网络Auto-FPN,本文方法的AP提高了1.7%.Fusion network is a representative module in object detection frameworks to fuse multi-scale features and improve detection accuracy.Previous works of designing fusion network architecture mainly focus on designing the topology of the fusion path to improve the performance of object detection.However,the required hardware resource overhead and the influence of feature selection and feature fusion operations on the detection performance are ignored.In this paper,we propose a feature fusion network named Attention-aware Fusion Network(AFN),which has a strong capacity of fusing multi-scale features for object detection.Through software and hardware cooperation,it can realize the automatic search of the neural network sensitive to hardware cost(parameter storage,calculation time,etc.),and realize the integrated optimization deployment from the three aspects of the fusion network’s characteristics,paths and operations.In this paper,we first summarize and propose three key factors that should be considered in the design of feature fusion network:fusion feature selection,fusion path and fusion mode.We also need to consider the hardware overhead of deploying the algorithm to the target platform.However,these design factors compose a huge design space that contains tremendous amounts of design choices.Thus,manually designing the optimal architecture of the fusion neck is very difficult.We employ neural network search(NAS)method to automatically design the feature fusion network.We propose three kind of search unit:feature search unit,fusion path search unit and fusion mode search unit.The feature search unit aims to search for the most appropriate input features for each scale instead of fixing from the top layer of each stage.The fusion path search unit takes all possible cross-scale fusing connections among groups as the search space and search for the optimal connection.the fusion mode search unit contains a variety of candidate fusion operations and decide operations to fuse features of multiple scales.Particularly,

关 键 词：目标检测 神经结构搜索 硬件开销 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]