Model efficiency has become increasingly important in computer vision. In this paper, we systematically study various neural network architecture design choices for object detection and propose several key optimizations to improve efficiency. First, we propose a weighted bi-directional feature pyramid network (BiFPN), which allows easy and fast multi-scale feature fusion; Second, we propose a compound scaling method that uniformly scales the resolution, depth, and width for all backbone, feature network, and box/class prediction networks at the same time. Based on these optimizations, we have developed a new family of object detectors, called EfficientDet, which consistently achieve an order-of-magnitude better efficiency than prior art across a wide spectrum of resource constraints. In particular, without bells and whistles, our EfficientDet-D7 achieves stateof-the-art 51.0 mAP on COCO dataset with 52M parameters and 326B FLOPS1 , being 4x smaller and using 9.3x fewer FLOPS yet still more accurate (+0.3% mAP) than the best previous detector.

5
下载
关闭预览

相关内容

谷歌大脑 Mingxing Tan、Ruoming Pang 和 Quoc V. Le 提出新架构 EfficientDet。EfficientDet检测器是单次检测器,非常类似于SSD和RetinaNet。骨干网络是ImageNet预训练的EfficientNet。把BiFPN用作特征网络,该网络从骨干网络获取3-7级{P3,P4,P5,P6,P7}特征,并反复应用自上而下和自下而上的双向特征融合。在广泛的资源限制下,这类模型的效率仍比之前最优模型高出一个数量级。具体来看,结构简洁只使用了 52M 参数、326B FLOPS 的 EfficientDet-D7 在 COCO 数据集上实现了当前最优的 51.0 mAP,准确率超越之前最优检测器(+0.3% mAP),其规模仅为之前最优检测器的 1/4,而后者的 FLOPS 更是 EfficientDet-D7 的 9.3 倍。

Detecting objects in aerial images is challenging for at least two reasons: (1) target objects like pedestrians are very small in pixels, making them hardly distinguished from surrounding background; and (2) targets are in general sparsely and non-uniformly distributed, making the detection very inefficient. In this paper, we address both issues inspired by observing that these targets are often clustered. In particular, we propose a Clustered Detection (ClusDet) network that unifies object clustering and detection in an end-to-end framework. The key components in ClusDet include a cluster proposal sub-network (CPNet), a scale estimation sub-network (ScaleNet), and a dedicated detection network (DetecNet). Given an input image, CPNet produces object cluster regions and ScaleNet estimates object scales for these regions. Then, each scale-normalized cluster region is fed into DetecNet for object detection. ClusDet has several advantages over previous solutions: (1) it greatly reduces the number of chips for final object detection and hence achieves high running time efficiency, (2) the cluster-based scale estimation is more accurate than previously used single-object based ones, hence effectively improves the detection for small objects, and (3) the final DetecNet is dedicated for clustered regions and implicitly models the prior context information so as to boost detection accuracy. The proposed method is tested on three popular aerial image datasets including VisDrone, UAVDT and DOTA. In all experiments, ClusDet achieves promising performance in comparison with state-of-the-art detectors. Code will be available in \url{https://github.com/fyangneil}.

0
4
下载
预览

Current state-of-the-art convolutional architectures for object detection are manually designed. Here we aim to learn a better architecture of feature pyramid network for object detection. We adopt Neural Architecture Search and discover a new feature pyramid architecture in a novel scalable search space covering all cross-scale connections. The discovered architecture, named NAS-FPN, consists of a combination of top-down and bottom-up connections to fuse features across scales. NAS-FPN, combined with various backbone models in the RetinaNet framework, achieves better accuracy and latency tradeoff compared to state-of-the-art object detection models. NAS-FPN improves mobile detection accuracy by 2 AP compared to state-of-the-art SSDLite with MobileNetV2 model in [32] and achieves 48.3 AP which surpasses Mask R-CNN [10] detection accuracy with less computation time.

0
6
下载
预览

It is a common paradigm in object detection frameworks to treat all samples equally and target at maximizing the performance on average. In this work, we revisit this paradigm through a careful study on how different samples contribute to the overall performance measured in terms of mAP. Our study suggests that the samples in each mini-batch are neither independent nor equally important, and therefore a better classifier on average does not necessarily mean higher mAP. Motivated by this study, we propose the notion of Prime Samples, those that play a key role in driving the detection performance. We further develop a simple yet effective sampling and learning strategy called PrIme Sample Attention (PISA) that directs the focus of the training process towards such samples. Our experiments demonstrate that it is often more effective to focus on prime samples than hard samples when training a detector. Particularly, On the MSCOCO dataset, PISA outperforms the random sampling baseline and hard mining schemes, e.g. OHEM and Focal Loss, consistently by more than 1% on both single-stage and two-stage detectors, with a strong backbone ResNeXt-101.

0
10
下载
预览

Although YOLOv2 approach is extremely fast on object detection; its backbone network has the low ability on feature extraction and fails to make full use of multi-scale local region features, which restricts the improvement of object detection accuracy. Therefore, this paper proposed a DC-SPP-YOLO (Dense Connection and Spatial Pyramid Pooling Based YOLO) approach for ameliorating the object detection accuracy of YOLOv2. Specifically, the dense connection of convolution layers is employed in the backbone network of YOLOv2 to strengthen the feature extraction and alleviate the vanishing-gradient problem. Moreover, an improved spatial pyramid pooling is introduced to pool and concatenate the multi-scale local region features, so that the network can learn the object features more comprehensively. The DC-SPP-YOLO model is established and trained based on a new loss function composed of mean square error and cross entropy, and the object detection is realized. Experiments demonstrate that the mAP (mean Average Precision) of DC-SPP-YOLO proposed on PASCAL VOC datasets and UA-DETRAC datasets is higher than that of YOLOv2; the object detection accuracy of DC-SPP-YOLO is superior to YOLOv2 by strengthening feature extraction and using the multi-scale local region features.

0
3
下载
预览

The task of detecting 3D objects in point cloud has a pivotal role in many real-world applications. However, 3D object detection performance is behind that of 2D object detection due to the lack of powerful 3D feature extraction methods. In order to address this issue, we propose to build a 3D backbone network to learn rich 3D feature maps by using sparse 3D CNN operations for 3D object detection in point cloud. The 3D backbone network can inherently learn 3D features from almost raw data without compressing point cloud into multiple 2D images and generate rich feature maps for object detection. The sparse 3D CNN takes full advantages of the sparsity in the 3D point cloud to accelerate computation and save memory, which makes the 3D backbone network achievable. Empirical experiments are conducted on the KITTI benchmark and results show that the proposed method can achieve state-of-the-art performance for 3D object detection.

0
7
下载
预览

Scale variation is one of the key challenges in object detection. In this work, we first present a controlled experiment to investigate the effect of receptive fields on the detection of different scale objects. Based on the findings from the exploration experiments, we propose a novel Trident Network (TridentNet) aiming to generate scale-specific feature maps with a uniform representational power. We construct a parallel multi-branch architecture in which each branch shares the same transformation parameters but with different receptive fields. Then, we propose a scale-aware training scheme to specialize each branch by sampling object instances of proper scales for training. As a bonus, a fast approximation version of TridentNet could achieve significant improvements without any additional parameters and computational cost. On the COCO dataset, our TridentNet with ResNet-101 backbone achieves state-of-the-art single-model results by obtaining an mAP of 48.4. Code will be made publicly available.

0
4
下载
预览

Lane mark detection is an important element in the road scene analysis for Advanced Driver Assistant System (ADAS). Limited by the onboard computing power, it is still a challenge to reduce system complexity and maintain high accuracy at the same time. In this paper, we propose a Lane Marking Detector (LMD) using a deep convolutional neural network to extract robust lane marking features. To improve its performance with a target of lower complexity, the dilated convolution is adopted. A shallower and thinner structure is designed to decrease the computational cost. Moreover, we also design post-processing algorithms to construct 3rd-order polynomial models to fit into the curved lanes. Our system shows promising results on the captured road scenes.

0
5
下载
预览

We'd like to share a simple tweak of Single Shot Multibox Detector (SSD) family of detectors, which is effective in reducing model size while maintaining the same quality. We share box predictors across all scales, and replace convolution between scales with max pooling. This has two advantages over vanilla SSD: (1) it avoids score miscalibration across scales; (2) the shared predictor sees the training data over all scales. Since we reduce the number of predictors to one, and trim all convolutions between them, model size is significantly smaller. We empirically show that these changes do not hurt model quality compared to vanilla SSD.

0
6
下载
预览

In this paper, we propose an efficient and fast object detector which can process hundreds of frames per second. To achieve this goal we investigate three main aspects of the object detection framework: network architecture, loss function and training data (labeled and unlabeled). In order to obtain compact network architecture, we introduce various improvements, based on recent work, to develop an architecture which is computationally light-weight and achieves a reasonable performance. To further improve the performance, while keeping the complexity same, we utilize distillation loss function. Using distillation loss we transfer the knowledge of a more accurate teacher network to proposed light-weight student network. We propose various innovations to make distillation efficient for the proposed one stage detector pipeline: objectness scaled distillation loss, feature map non-maximal suppression and a single unified distillation loss function for detection. Finally, building upon the distillation loss, we explore how much can we push the performance by utilizing the unlabeled data. We train our model with unlabeled data using the soft labels of the teacher network. Our final network consists of 10x fewer parameters than the VGG based object detection network and it achieves a speed of more than 200 FPS and proposed changes improve the detection accuracy by 14 mAP over the baseline on Pascal dataset.

0
5
下载
预览

Object detection is a major challenge in computer vision, involving both object classification and object localization within a scene. While deep neural networks have been shown in recent years to yield very powerful techniques for tackling the challenge of object detection, one of the biggest challenges with enabling such object detection networks for widespread deployment on embedded devices is high computational and memory requirements. Recently, there has been an increasing focus in exploring small deep neural network architectures for object detection that are more suitable for embedded devices, such as Tiny YOLO and SqueezeDet. Inspired by the efficiency of the Fire microarchitecture introduced in SqueezeNet and the object detection performance of the single-shot detection macroarchitecture introduced in SSD, this paper introduces Tiny SSD, a single-shot detection deep convolutional neural network for real-time embedded object detection that is composed of a highly optimized, non-uniform Fire sub-network stack and a non-uniform sub-network stack of highly optimized SSD-based auxiliary convolutional feature layers designed specifically to minimize model size while maintaining object detection performance. The resulting Tiny SSD possess a model size of 2.3MB (~26X smaller than Tiny YOLO) while still achieving an mAP of 61.3% on VOC 2007 (~4.2% higher than Tiny YOLO). These experimental results show that very small deep neural network architectures can be designed for real-time object detection that are well-suited for embedded scenarios.

0
7
下载
预览
小贴士
相关论文
Clustered Object Detection in Aerial Images
Fan Yang,Heng Fan,Peng Chu,Erik Blasch,Haibin Ling
4+阅读 · 2019年8月27日
Golnaz Ghiasi,Tsung-Yi Lin,Ruoming Pang,Quoc V. Le
6+阅读 · 2019年4月16日
Yuhang Cao,Kai Chen,Chen Change Loy,Dahua Lin
10+阅读 · 2019年4月9日
Xuesong Li,Jose E Guivant,Ngaiming Kwok,Yongzhi Xu
7+阅读 · 2019年1月24日
Scale-Aware Trident Networks for Object Detection
Yanghao Li,Yuntao Chen,Naiyan Wang,Zhaoxiang Zhang
4+阅读 · 2019年1月7日
Efficient Road Lane Marking Detection with Deep Learning
Ping-Rong Chen,Shao-Yuan Lo,Hsueh-Ming Hang,Sheng-Wei Chan,Jing-Jhih Lin
5+阅读 · 2018年9月11日
Pengchong Jin,Vivek Rathod,Xiangxin Zhu
6+阅读 · 2018年7月9日
Rakesh Mehta,Cemalettin Ozturk
5+阅读 · 2018年5月16日
Alexander Wong,Mohammad Javad Shafiee,Francis Li,Brendan Chwyl
7+阅读 · 2018年2月19日
相关VIP内容
[综述]深度学习下的场景文本检测与识别
专知会员服务
41+阅读 · 2019年10月10日
相关资讯
Ray RLlib: Scalable 降龙十八掌
CreateAMind
5+阅读 · 2018年12月28日
Focal Loss for Dense Object Detection
统计学习与视觉计算组
11+阅读 · 2018年3月15日
ResNet, AlexNet, VGG, Inception:各种卷积网络架构的理解
全球人工智能
15+阅读 · 2017年12月17日
【推荐】ResNet, AlexNet, VGG, Inception:各种卷积网络架构的理解
机器学习研究会
17+阅读 · 2017年12月17日
动手写机器学习算法:异常检测 Anomaly Detection
七月在线实验室
9+阅读 · 2017年12月8日
Capsule Networks解析
机器学习研究会
10+阅读 · 2017年11月12日
【推荐】YOLO实时目标检测(6fps)
机器学习研究会
16+阅读 · 2017年11月5日
【推荐】深度学习目标检测全面综述
机器学习研究会
17+阅读 · 2017年9月13日
【推荐】深度学习目标检测概览
机器学习研究会
9+阅读 · 2017年9月1日
Top