Object Detection with Low Capacity GPU Systems Using Improved Faster R-CNN

Abstract

Object detection in remote sensing images has been frequently used in a wide range of areas such as land planning, city monitoring, traffic monitoring, and agricultural applications. It is essential in the field of aerial and satellite image analysis but it is also a challenge. To overcome this challenging problem, there are many object detection models using convolutional neural networks (CNN). The deformable convolutional structure has been introduced to eliminate the disadvantage of the fixed grid structure of the convolutional neural networks. In this study, a multi-scale Faster R-CNN method based on deformable convolution is proposed for single/low graphics processing unit (GPU) systems. Weight standardization (WS) is used instead of batch normalization (BN) to make the proposed model more efficient for a small batch size (1 img/per GPU) on single GPU systems. Experiments were conducted on the publicly available 10-class geospatial object detection (NWPU-VHR 10) dataset to evaluate the object detection performance of the proposed model. Experiment results show that our model achieved a 92.3 mAP. This is a 1.7% mAP increase when compared to the best results in the models using the same dataset.

Weighted Ensemble Object Detection with Optimized Coefficients for Remote Sensing Images

“…The results given in Table 3 reveal that the proposed model provides better results than those in other studies. Although our method ensures high performance, we observed a discrepancy in the detection accuracy for the same category of object (e.g., storage tank) as in the DODN [12] and Improved Faster R-CNN [13] models. This is mainly because of the imbalance among the classes in the NWPU-VHR10 database.…”

“…The present model is based on the latest Faster R-CNN (Figure 3) [26], a state-of-the-art object detection system. The deformable convolution [27] technique is used to overcome the weaknesses of the regular convolution structure used in the Faster R-CNN model for detecting small and mixed objects in remote sensing images [13]. With the Feature Pyramid Network (FPN) [28] technique, the high-resolution features in the shallow layers of the remote sensing images are transferred to the network.…”

“…Wang et al [12] proposed a new region proposal network that is anchor-free and sliding-window-free to develop a two-stage object detection network. Körez et al [13] improved the Faster R-CNN algorithm via the deformable convolution and the weight standardization techniques for remote sensing object detection using low-capacity Graphic Processing Units (GPUs). Qui et al [14] proposed a novel end-to-end Adaptive Aspect Ratio Multiscale Network (A 2 RMNet), which is a multiscale feature gate fusion network used to adaptively integrate the multiscale features of objects.…”

Mixed YOLOv3-LITE: A Lightweight Real-Time Object Detection Method

“…In Equations (4) and (5), D represents the number of layers of the CNN, i.e., the depth of the network; l represents the lth convolution layer of the CNN; M l represents the side length of the output feature map for the lth convolution layer; K represents the side length of each convolution kernel; C l−1 represents the number of input channels of the lth convolution layer, i.e., the number of output channels of the (l − 1)th convolution layer; and C l represents the number of output channels of the lth convolution layer, i.e., the number of convolution kernels of this layer.…”

“…Feature extraction and classification of original images can be conducted via multi-layer convolution operations, and the position of an object in an image can be predicted using boundary boxes, providing the capability of visual understanding. The results of these studies can be widely applied in facial recognition [1], attitude prediction [2], video surveillance, and a variety of other intelligent applications [3][4][5].…”