English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The record-breaking achievements of deep neural networks (DNNs) in imageclassification and detection tasks resulted in a surge of new computer visionapplications during the past years. However, their computational complexity isrestricting their deployment to powerful stationary or complex dedicatedprocessing hardware, limiting their use in smart edge processing applications.We propose an automated deployment framework for DNN acceleration at the edgeon field-programmable gate array (FPGA)-based cameras. The frameworkautomatically converts an arbitrary-sized and quantized trained network into anefficient streaming-processing IP block that is instantiated within a genericadapter block in the FPGA. In contrast to prior work, the accelerator is purelylogic and thus supports end-to-end processing on FPGAs without on-chipmicroprocessors. Our mapping tool features automatic translation from a trainedCaffe network, arbitrary layer-wise fixed-point precision for both weights andactivations, an efficient XNOR implementation for fully binary layers as wellas a balancing mechanism for effective allocation of computational resources inthe streaming dataflow. To present the performance of the system we employ thistool to implement two CNN edge processing networks on an FPGA-based high-speedcamera with various precision settings showing computational throughputs of upto 337GOPS in low-latency streaming mode (no batching), running entirely on thecamera.

NN2CAM: Automated Neural Network Mapping for Multi-Precision Edge Processing on FPGA-Based Cameras