Evaluating pruned object detection networks for real-time robot vision

O'Keeffe, Simon; Villing, Rudi

doi:10.1109/icarsc.2018.8374166

Cited by 11 publications

(33 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The absolute weight sum technique was successful in reducing the computational cost for VGG-16 [19] by up to 34% and ResNet-110 [20] by up to 38% with no significant loss in accuracy [17]. We have previously shown that pruning methods based on Taylor expansion did not outperform an approach based on the absolute weight sum [7].…”

Section: Convolutionalmentioning

confidence: 99%

“…In the literature, pruning has been evaluated for image classification problems but not for object detection (as far as the authors are aware) except for our own work [7]. There is reason to suspect object detection networks might be less amenable to pruning as more information is carried through the network to the output layer.…”

Section: Convolutionalmentioning

confidence: 99%

“…There is reason to suspect object detection networks might be less amenable to pruning as more information is carried through the network to the output layer. We have shown that a standard pruning procedure could be effectively applied to an object detection network trained on a small dataset [7] but it is not yet clear whether pruning can be applied with equal success to an object detection network trained on a more challenging dataset such as PASCAL VOC.…”

Section: Convolutionalmentioning

confidence: 99%

“…In this work, we propose a new Extended Pruning procedure in which there are multiple pruning phases and extended fine-tuning is performed at intervals to maintain network accuracy for longer. The motivation for our procedure is that a single long pruning phase may reduce network accuracy more than can be recovered by extended fine-tuning just once at the end [7], whereas by pruning with multiple extended finetuning phases to recover lost accuracy we can prune networks to a greater extent than would otherwise be feasible.…”

Section: Introductionmentioning

confidence: 99%

“…Standard pruning has typically been performed on classification networks only and has only been evaluated by us on the potentially more challenging case of an object detection network [7]. Therefore in this work we evaluate our new Extended Pruning procedure on two object detection networks: a network trained for a small custom dataset and a network trained on the more challenging PASCAL VOC dataset [8] which we expect might not have as much redundancy.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Evaluating Extended Pruning on Object Detection Neural Networks

O'Keeffe

Villing

2018

2018 29th Irish Signals and Systems Conference (ISSC)

Self Cite

View full text Add to dashboard Cite

CNNs are the state-of-the-art for many computer vision problems, including object detection. However, reducing the computational complexity of a CNN is a key prerequisite to deploying state-of-the-art deep learning networks in many low power embedded real-time robotic applications. Pruning has been shown to be an effective method to reduce the computational complexity of a Convolutional Neural Network (CNN) while maintaining accuracy. In the literature, accuracy lost through pruning is recovered with extended fine-tuning of the pruned network at the end of the pruning procedure, but further pruning is not conducted after extended fine-tuning. In this work we modify the pruning procedure to incorporate extended fine-tuning at intervals during the procedure to maintain network accuracy while pruning further than would otherwise be possible. We evaluate this procedure on a small scale custom object detection dataset and the more challenging standard PASCAL VOC dataset. On the former the new procedure achieves a 19.6× reduction in FLOPS for a drop of only 0.4% mean Average Precision (mAP) while the latter achieves only a 1.8× reduction in FLOPS for a drop of 0.8% mAP. The results indicate differing levels of parameter redundancy in the initial networks.

show abstract

Section: Convolutionalmentioning

confidence: 99%

Section: Convolutionalmentioning

confidence: 99%

Section: Convolutionalmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Evaluating Extended Pruning on Object Detection Neural Networks

O'Keeffe

Villing

2018

2018 29th Irish Signals and Systems Conference (ISSC)

Self Cite

View full text Add to dashboard Cite

show abstract

YOLOpeds: efficient real‐time single‐shot pedestrian detection for smart camera applications

Kyrkou

2020

IET Computer Vision

View full text Add to dashboard Cite

Deep‐learning‐based pedestrian detectors can enhance the capabilities of smart camera systems in a wide spectrum of machine vision applications including video surveillance, autonomous driving, robots and drones, smart factory, and health monitoring. However, such complex paradigms do not scale easily and are not traditionally implemented in resource‐constrained smart cameras for on‐device processing which offers significant advantages in situations when real‐time monitoring and privacy are vital. This work addresses the challenge of achieving a good trade‐off between accuracy and speed for efficient deep‐learning‐based pedestrian detection in smart camera applications. The contributions of this work are the following: 1) a computationally efficient architecture based on separable convolutions that integrates dense connections across layers and multi‐scale feature fusion to improve representational capacity while decreasing the number of parameters and operations, 2) a more elaborate loss function for improved localization, 3) and an anchor‐less approach for detection. The proposed approach referred to as YOLOpeds is evaluated using the PETS2009 surveillance dataset on 320 × 320 images. A real‐system implementation is presented using the Jetson TX2 embedded platform. YOLOpeds provides real‐time sustained operation of over 30 frames per second with detection rates in the range of 86% outperforming existing deep learning models.

show abstract

PG-YOLO: A Novel Lightweight Object Detection Method for Edge Devices in Industrial Internet of Things

et al. 2022

View full text Add to dashboard Cite

With the rapid development of Industrial Internet of Things (IIoT) technology, video surveillance devices and video data in IIoT environments are massively growing and increasingly important. Deploying rapidly evolving deep learning-based object detection algorithms in IIoT can improve the efficiency of video data utilization and increase the automation and intelligence of the IIoT. Facing the transmission latency problem of massive video data, the algorithms are better deployed in edge devices. However, due to the large size and high computing power requirements of existing object detection algorithms, it is difficult to deploy them on edge devices with limited performance. Therefore, to solve this problem, we propose a smaller and less computationally intensive object detection algorithm, PG-YOLO. We replaced all the convolutional modules of YOLOV5 with low computational ghost modules and streamlined the backbone network to improve the network performance. Then, we propose a better pruning algorithm to compress the network, and finally improve the accuracy by distillation to obtain PG-YOLO. PG-YOLO is smaller and faster and can be deployed in edge devices. In addition, we conducted experimental validation on the SHWD dataset. The experimental results show that PG-YOLO can compress the volume of the model by 9 times compared with YOLOv5s, and the compressed detection accuracy reaches 0.934, with only 0.1% loss of accuracy. Also, the time required for inference is reduced by 32.7%, and the inference speed is improved by 10 FPS. Compared with other object detection algorithms, PG-YOLO also has advantages.INDEX TERMS deep learning, object detection, edge device, lightweight model, PG-YOLO.

show abstract

Evaluating pruned object detection networks for real-time robot vision

Cited by 11 publications

References 14 publications

Evaluating Extended Pruning on Object Detection Neural Networks

Evaluating Extended Pruning on Object Detection Neural Networks

YOLOpeds: efficient real‐time single‐shot pedestrian detection for smart camera applications

PG-YOLO: A Novel Lightweight Object Detection Method for Edge Devices in Industrial Internet of Things

Contact Info

Product

Resources

About