Review of prominent strategies for mapping CNNs onto embedded systems

Arredondo-Velázquez, Moisés; Díaz-Carmona, Javier; Barranco-Gutiérrez, Alejandro I.; Torres-Huitzil, César

doi:10.1109/tla.2020.9082927

Cited by 13 publications

(4 citation statements)

References 89 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A deep neural network needs a huge amount of computation, which is usually deployed to edge devices by heterogeneous computing units [ 22 , 23 , 24 ]. These heterogeneous computing units are still in the research stage in China.…”

Section: Introductionmentioning

confidence: 99%

Detection of Missing Insulator Caps Based on Machine Learning and Morphological Detection

Zhang

Chen

Huang

2023

Sensors

View full text Add to dashboard Cite

Missing insulator caps are the key focus of transmission line inspection work. Insulators with a missing cap will experience decreased insulation and mechanical strength and cause transmission line safety accidents. As missing insulator caps often occur in glass and porcelain insulators, this paper proposes a detection method for missing insulator caps in these materials. First, according to the grayscale and color characteristics of these insulators, similar characteristic regions of the insulators are extracted from inspection images, and candidate boxes are generated based on these characteristic regions. Second, the images captured by these boxes are input into the classifier composed of SVM (Support Vector Machine) to identify and locate the insulators. The accuracy, recall and average accuracy of the classifier are all higher than 90%. Finally, this paper proposes a processing method based on the insulator morphology to determine whether an insulator cap is missing. The proposed method can also detect the number of remaining insulators, which can help power supply enterprises to evaluate the degree of insulator damage.

show abstract

Section: Introductionmentioning

confidence: 99%

Detection of Missing Insulator Caps Based on Machine Learning and Morphological Detection

Zhang

Chen

Huang

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…Overall, the highest accuracy was reached by introducing novel network components and irregular connectivity between layers that differ from the conventional layer types [21,22]. To deal with a wide range of CNN configurations, algorithm-level solutions [23] used in CPU-or GPU-based platforms have been employed either for their computing capabilities or the several deep learning frameworks (Caffe, Theano, Keras or Tensorflow [24,25]) that enable sharing trained network parameters for new custom models. Nevertheless, these approaches are still unsuitable for resource-constrained hardware, particularly in GPU devices, where power consumption makes implementing batteries-based systems unfeasible.…”

Section: Introductionmentioning

confidence: 99%

Flexible Convolver for Convolutional Neural Networks Deployment onto Hardware-Oriented Applications

et al. 2022

Self Cite

View full text Add to dashboard Cite

This paper introduces a flexible convolver capable of adapting to the different convolution layer configurations of state-of-the-art Convolution Neural Networks (CNNs). The use of two proposed programmable components achieves this adaptability. A Programmable Line Buffer (PLB) based on Programmable Shift Registers (PSRs) allows the generation of the required convolution masks required for each processed CNN layer. The convolution layer computing is performed through a proposed programmable systolic array configured according to the target device resources. In order to maximize the device resource usage and to achieve a shortened processing time, the filter, data, and loop parallelisms are leveraged. These characteristics allow the described architecture to be scalable and implemented on any FPGA device targeting different applications. The convolver description was written in VHDL using the Intel Cyclone V 5CSXFC6D6F31C6N device as a reference. The experimental results show that the proposed computing method allows the processing of any CNN without requiring special adaptation for a specific application since the standard convolution algorithm is used. The proposed flexible convolver achieves competitive performance compared with those reported in related works.

show abstract

“…It is especially appropriate for mass production. (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18) Deep learning (DL) has been adopted in AOI, which is a data-driven algorithm that learns how to classify objects from training data. DL learns to detect defective beans and foreign objects from a database of numerous images of defective and good-quality beans.…”

Section: Introductionmentioning

confidence: 99%

“…A CNN shows good performance in image classification or recognition but has high computational complexity. As a CNN requires large storage space and computing resources, (9) it cannot be used on mobile devices or embedded platforms. This problem must be solved to implement a smart coffee bean inspection system.…”

Section: Introductionmentioning

confidence: 99%

Deep Convolutional Neural Network for Coffee Bean Inspection

Wang¹,

Tseng²,

Chen³

et al. 2021

Sensors and Materials

View full text Add to dashboard Cite

Coffee is one of the most popular drinks in the world. It contains antioxidants and healthpromoting nutrients that can boost one's energy and focus. However, defective beans mixed in with raw beans can easily affect the flavor and even be harmful to human health. The traditional human visual inspection of defective beans is extremely laborious and time-consuming and may result in low-quality coffee due to worker stress and fatigue. We propose a lightweight and explainable intelligent coffee bean quality inspection system that uses deep learning (DL) and computer vision (CV) technologies to assist operators in detecting defects, including mold, fermentation, insect bites, and crushed beans. We use knowledge distillation (KD) to achieve model compression. The basic explainable convolutional neural network (CNN) model is established using the explainable AI (XAI) method. The implemented system has a high identification rate, low complexity, and low power consumption, and can explain the judgment criteria of the complex classification model.

show abstract

Review of prominent strategies for mapping CNNs onto embedded systems

Cited by 13 publications

References 89 publications

Detection of Missing Insulator Caps Based on Machine Learning and Morphological Detection

Detection of Missing Insulator Caps Based on Machine Learning and Morphological Detection

Flexible Convolver for Convolutional Neural Networks Deployment onto Hardware-Oriented Applications

Deep Convolutional Neural Network for Coffee Bean Inspection

Contact Info

Product

Resources

About