Object Detection at 200 Frames per Second

Mehta, Rakesh; Öztürk, Cemalettin

doi:10.1007/978-3-030-11021-5_41

Cited by 53 publications

(27 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our DRFB module used 1×1 convolution to increase nonlinearity and depth. This minimizes the amount of computation increases and improves the capacity of the structure [50]. Instead of using 3×3 convolutions, 1×3 and 3×1 convolutions were used to reduce computational complexity with nonlinearity increments.…”

Section: Proposed Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Real-Time Vehicle-Detection Method in Bird-View Unmanned-Aerial-Vehicle Imagery

Han

Yoo

Kwon

2019

Sensors

View full text Add to dashboard Cite

Vehicle detection is an important research area that provides background information for the diversity of unmanned-aerial-vehicle (UAV) applications. In this paper, we propose a vehicle-detection method using a convolutional-neural-network (CNN)-based object detector. We design our method, DRFBNet300, with a Deeper Receptive Field Block (DRFB) module that enhances the expressiveness of feature maps to detect small objects in the UAV imagery. We also propose the UAV-cars dataset that includes the composition and angular distortion of vehicles in UAV imagery to train our DRFBNet300. Lastly, we propose a Split Image Processing (SIP) method to improve the accuracy of the detection model. Our DRFBNet300 achieves 21 mAP with 45 FPS in the MS COCO metric, which is the highest score compared to other lightweight single-stage methods running in real time. In addition, DRFBNet300, trained on the UAV-cars dataset, obtains the highest AP score at altitudes of 20–50 m. The gap of accuracy improvement by applying the SIP method became larger when the altitude increases. The DRFBNet300 trained on the UAV-cars dataset with SIP method operates at 33 FPS, enabling real-time vehicle detection.

show abstract

Section: Proposed Methodsmentioning

confidence: 99%

“…To train the general object-detection model, most studies used datasets such as MS COCO [23] or PASCAL VOC [13,14,15,16,17,18,20,24,40,50]. Each dataset has 81 and 21 labels, including backgrounds, and labels such as frisbee, hot dog, and potted plant.…”

Section: Proposed Methodsmentioning

confidence: 99%

Real-Time Vehicle-Detection Method in Bird-View Unmanned-Aerial-Vehicle Imagery

Han

Yoo

Kwon

2019

Sensors

View full text Add to dashboard Cite

show abstract

“…Mehta and Ozturk [ 41 ] proposed F-YOLO based on Tiny-YOLO. Generally, a deeper and wider network structure, results in a better detection effect, but the corresponding calculation of parameters will also increase, resulting in slow network training.…”

Section: Algorithm Principle and Df-tiny-yolo Networkmentioning

confidence: 99%

“…Practical problems such as complexity and variations in the veins of apple leaves and difficulties with disease identification can be overcome by the rapid and effective automated detection of apple leaf diseases. Therefore, we created the DF-Tiny-YOLO model to, an detect apple leaf diseases based on regression with the concepts of DenseNet [ 40 ] and F-YOLO [ 41 ]. We then optimized the model considering rapid, accurate detection.…”

Section: Introductionmentioning

confidence: 99%

A method of detecting apple leaf diseases based on improved convolutional neural network

2022

PLoS ONE

View full text Add to dashboard Cite

Apple tree diseases have perplexed orchard farmers for several years. At present, numerous studies have investigated deep learning for fruit and vegetable crop disease detection. Because of the complexity and variety of apple leaf veins and the difficulty in judging similar diseases, a new target detection model of apple leaf diseases DF-Tiny-YOLO, based on deep learning, is proposed to realize faster and more effective automatic detection of apple leaf diseases. Four common apple leaf diseases, including 1,404 images, were selected for data modeling and method evaluation, and made three main improvements. Feature reuse was combined with the DenseNet densely connected network and further realized to reduce the disappearance of the deep gradient, thus strengthening feature propagation and improving detection accuracy. We introduced Resize and Re-organization (Reorg) and conducted convolution kernel compression to reduce the calculation parameters of the model, improve the operating detection speed, and allow feature stacking to achieve feature fusion. The network terminal uses convolution kernels of 1 × 1, 1 × 1, and 3 × 3, in turn, to realize the dimensionality reduction of features and increase network depth without increasing computational complexity, thus further improving the detection accuracy. The results showed that the mean average precision (mAP) and average intersection over union (IoU) of the DF-Tiny-YOLO model were 99.99% and 90.88%, respectively, and the detection speed reached 280 FPS. Compared with the Tiny-YOLO and YOLOv2 network models, the new method proposed in this paper significantly improves the detection performance. It can also detect apple leaf diseases quickly and effectively.

show abstract

“…In Mehta and Ozturk work [7], the authors used Knowledge Distillation for object detection. The student network is the Tiny-YOLO network, and the YOLOv2 network is the teacher network.…”

Section: Introductionmentioning

confidence: 99%

Squeezed Deep 6DoF Object Detection Using Knowledge Distillation

Felix,

Rodrigues,

Macêdo

et al. 2020

Preprint

View full text Add to dashboard Cite

The detection of objects considering a 6DoF pose is common requisite to build virtual and augmented reality applications. It is usually a complex task witch requires realtime processing and high precision results for an adequate user experience. Recently, different deep learning techniques have been proposed to detect objects in 6DoF in RGB images but they rely on high complexity networks, requiring a computational power that prevents them to work on mobile devices. In this paper, we propose an approach to reduce the complexity of 6DoF detection networks while maintaining accuracy. We used Knowledge Distillation to teach portables Convolutional Neural Networks (CNN) to learn from a realtime 6DoF detection CNN. The proposed method allows realtime applications using only RGB images while decreasing the hardware requirements. We used the LINEMOD dataset to evaluate the proposed method and the experimental results show that the proposed method reduces the memory requirement almost 99% in comparison to the original architecture reducing half the accuracy in one of the metrics. Code is available at https://github.com/heitorcfelix/singleshot6Dpose.

show abstract

Object Detection at 200 Frames per Second

Cited by 53 publications

References 39 publications

Real-Time Vehicle-Detection Method in Bird-View Unmanned-Aerial-Vehicle Imagery

Real-Time Vehicle-Detection Method in Bird-View Unmanned-Aerial-Vehicle Imagery

A method of detecting apple leaf diseases based on improved convolutional neural network

Squeezed Deep 6DoF Object Detection Using Knowledge Distillation

Contact Info

Product

Resources

About