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

IoT devices suffer from resource limitations, such as processor, RAM, anddisc storage. These limitations become more evident when handling demandingapplications, such as deep learning, well-known for their heavy computationalrequirements. A case in point is robot pose estimation, an application thatpredicts the critical points of the desired image object. One way to mitigateprocessing and storage problems is compressing that deep learning application.This paper proposes a new CNN for the pose estimation while applying thecompression techniques of pruning and quantization to reduce his demands andimprove the response time. While the pruning process reduces the total numberof parameters required for inference, quantization decreases the precision ofthe floating-point. We run the approach using a pose estimation task for arobotic arm and compare the results in a high-end device and a constraineddevice. As metrics, we consider the number of Floating-point Operations PerSecond(FLOPS), the total of mathematical computations, the calculation ofparameters, the inference time, and the number of video frames processed persecond. In addition, we undertake a qualitative evaluation where we compare theoutput image predicted for each pruned network with the corresponding originalone. We reduce the originally proposed network to a 70% pruning rate, implyingan 88.86% reduction in parameters, 94.45% reduction in FLOPS, and for the discstorage, we reduced the requirement in 70% while increasing error by a mere$1\%$. With regard input image processing, this metric increases from 11.71 FPSto 41.9 FPS for the Desktop case. When using the constrained device, imageprocessing augmented from 2.86 FPS to 10.04 FPS. The higher processing rate ofimage frames achieved by the proposed approach allows a much shorter responsetime.

FCN-Pose: A Pruned and Quantized CNN for Robot Pose Estimation for Constrained Devices