Fast and Lightweight Object Detection Network: Detection and Recognition on Resource Constrained Devices

Oliveira, Bernardo Augusto Godinho de; Ferreira, Flavia Magalhaes Freitas; Martins, Carlos Augusto Paiva da Silva

doi:10.1109/access.2018.2801813

Cited by 24 publications

(13 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For many inferencing tasks, general-purpose processors are not required, so researchers found that GPUs provide full resolution of floating-point arithmetic therefore network pruning uses post-training DNN analysis to eliminate connections between neurons that have little or no effect on performance [10]. The needless calculations and, as an additional benefit, significantly reduce the energy consumed by memory accesses due to the deletion of their connections.…”

Section: Machine Learning In Embedded Systemsmentioning

confidence: 99%

“…One of it's state-of -the-art criteria for tracking of an object is the visualization challenge [8] as well as the results of past years have indeed been focused on both deep features [11] and deep learning techniques [9]. The developing and challenging of algorithms and CNNs endlessly to become more accurate and quicker but are typically evaluated to benchmark databases [10] on strong hardware platforms. At the other side, designers and researchers are looking to diversify the hardware available options to provide all resources required for implementation to improve efficiency of such challenging networks.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Heuristics of Machine Learning on Embedded Intelligence Device for Multiple Object Detection

C¹

2020

IJETER

View full text Add to dashboard Cite

The application of Machine Learning algorithms in the embedded intelligence device is still a challenge to many researchers. This can be achieved by implementing Machine Learning algorithms in devices like computers or mini-computers and later this can be extended to high performance embedded intelligent boards. In this project the primary focus will be on how the Machine learning algorithms provide the data processing in the given data sets to provide an optimal solution. In this way the motion of objects pertaining to an area can be easily tracked and identified. If any intrusion occurs then it can be easily detected without any further delay. For this process to happen the various aspects are a cloud based server which the embedded intelligence device is connected, a suitable and efficient Machine learning algorithm and lastly inclusion of the required data sets for the detection or we can use the data sets that are already present in the open cloud server.

show abstract

Section: Machine Learning In Embedded Systemsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Heuristics of Machine Learning on Embedded Intelligence Device for Multiple Object Detection

C¹

2020

IJETER

View full text Add to dashboard Cite

show abstract

“…Oliveira et al [13] proposed a Fast and Light Weight Object Detection (FLODNet) model based on Convolutionalneural network which is faster than CNN model and can execute in CPU within few seconds. The FLODNet is a shallow convolutional neural network with 10 layers and fixed kernel size.…”

Section: Related Workmentioning

confidence: 99%

Object Recognition in Image Sequence using Heuristic Convolution Neural Network

2019

IJEAT

View full text Add to dashboard Cite

The human visual system can make a distinction of tiger from cat very easily without taking any efforts. But in case of a computer system, it is a very complicated job. Identifying and differentiating task has to deal with many challenges but the human brain makes it effortless. Self learning or heuristic techniques are most relevant in this area. The recognition task is to search for the particular object of same shape, color and texture and so on, of the trained objects and match with input. The geometrical distinction such as zoom in, zoom out, rotation etc result in poor performance. This paper uses convolution neural network models Alexnet and VGGNet on object recognition problems which are added with novel heuristic method. We have used CIFAR-10 dataset. The performance and computation speeds are found efficient.

show abstract

“…Real-time object detection techniques have been applied to a variety of computer vision areas [1,2], such as object classification or object segmentation. Since it is mainly operated on the constrained environments, input images obtained from those environments can be deteriorated by camera noises or compression artifacts [3][4][5]. In particular, it is hard to detect objects from the images with low quality.…”

Section: Introductionmentioning

confidence: 99%

Enhanced Single Image Super Resolution Method Using Lightweight Multi-Scale Channel Dense Network

Lee

Jun

Kim

et al. 2021

Sensors

View full text Add to dashboard Cite

Super resolution (SR) enables to generate a high-resolution (HR) image from one or more low-resolution (LR) images. Since a variety of CNN models have been recently studied in the areas of computer vision, these approaches have been combined with SR in order to provide higher image restoration. In this paper, we propose a lightweight CNN-based SR method, named multi-scale channel dense network (MCDN). In order to design the proposed network, we extracted the training images from the DIVerse 2K (DIV2K) dataset and investigated the trade-off between the SR accuracy and the network complexity. The experimental results show that the proposed method can significantly reduce the network complexity, such as the number of network parameters and total memory capacity, while maintaining slightly better or similar perceptual quality compared to the previous methods.

show abstract

Fast and Lightweight Object Detection Network: Detection and Recognition on Resource Constrained Devices

Cited by 24 publications

References 35 publications

Heuristics of Machine Learning on Embedded Intelligence Device for Multiple Object Detection

Heuristics of Machine Learning on Embedded Intelligence Device for Multiple Object Detection

Object Recognition in Image Sequence using Heuristic Convolution Neural Network

Enhanced Single Image Super Resolution Method Using Lightweight Multi-Scale Channel Dense Network

Contact Info

Product

Resources

About