Patent
Embedded system-oriented neural network mapping method and device
Yu Bo,Liu Shaoshan +1 more
- 24 Apr 2018
3
TL;DR: In this article, an embedded system-oriented neural network mapping method and device is presented, which comprises the following steps of: establishing a neural network structure and obtaining parameters of the neural network; generating a data flow chart for describing the neural networks according to the network structure; obtaining a deep learning calculation unit library; and realizing the data flowchart of neural network by utilizing the deep learning computation unit library so as to obtain a realization program of the network.
read more
Abstract: The invention discloses an embedded system-oriented neural network mapping method and device. The method comprises the following steps of: establishing a neural network structure and obtaining parameters of the neural network; generating a data flow chart for describing the neural network according to the neural network structure and the parameter of the neural network; obtaining a deep learning calculation unit library and realizing the data flow chart of the neural network by utilizing the deep learning calculation unit library so as to obtain a realization program of the neural network. According to the method and device, the neural network can be mapped into the realization program applied to an embedded system by taking the deep learning calculation unit, so that the realization speed, on an embedded platform, of the neural network can be improved and performance of the embedded platform can be maximized.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Patent
Visual-inertial positional awareness for autonomous and non-autonomous tracking
Zhe Zhang,Grace Tsai,Shaoshan Liu +2 more
- 15 Jun 2018
TL;DR: In this paper, positional awareness techniques employing visual-inertial sensory data gathering and analysis hardware with reference to specific example implementations implement improvements in the use of sensors, techniques and hardware design that can enable specific embodiments to provide positional awareness to machines with improved speed and accuracy.
65
Patent
Target object detection method and device, electronic equipment and storage medium
Bo Li,Wu Wei,Fangyi Zhang +2 more
- 07 May 2019
TL;DR: In this paper, a target object detection method and device, electronic equipment and a storage medium is described, which comprises the steps: extracting the characteristics of a reference frame and a detection frame through a neural network, where the image size of the reference frame is smaller than that of the detection frame.
2
Patent
Deep learning model generation method and apparatus, device, and storage medium
Tan Zhipeng,Liu Yaoyong,Jiang Yi +2 more
- 01 Apr 2021
TL;DR: In this paper, a deep learning model generation method and apparatus, a device, and a storage medium are described, and the method comprises: generating a first source file according to a model file of deep learning, the model file comprising a weight matrix in the deep learning models, obtaining a second source file corresponding to the deep models, and compiling the first source files and the second source files to generate a target model.
References
OpenCL caffe: Accelerating and enabling a cross platform machine learning framework
Junli Gu,Yibing Liu,Yuan Gao,Maohua Zhu +3 more
- 19 Apr 2016
TL;DR: This paper presents OpenCL™ caffe, which targets in transforming the popular CUDA based framework caffe into open standard OpenCL backend to enable a heterogeneous platform compatible DNN framework and achieve competitive performance based on OpenCL tool chain.
37
LCDet: Low-Complexity Fully-Convolutional Neural Networks for Object Detection in Embedded Systems
Subarna Tripathi,Gokce Dane,Byeongkeun Kang,Vasudev Bhaskaran,Truong Q. Nguyen +4 more
- 01 Jul 2017
TL;DR: Li et al. as discussed by the authors proposed an end-to-end TensorFlow-based model for generic object detection that aims to work in em- bedded systems, which works by a single forward pass through the network and employs 8-bit quantization on the learned weights.
Patent
Method for optimizing artificial neural network
Yao Song
- 10 Oct 2017
TL;DR: In this article, a CNN accelerator is provided for image-net large-scale classification based on the embedded FPGA platform, which can be applied to imagenet large scale image classification.
35
Patent
Single image rain removal method based on convolutional neural network
Ding Xinghao,Fu Xueyang,Chen Liqin +2 more
- 07 Dec 2016
TL;DR: In this paper, a single image rain removal method based on a convolutional neural network and related to image processing is proposed. But the method is not suitable for real-time image processing.
13