Abstract
Fortheproblemsthat inthecollaborativeintelligenceframework theintermediate featuredataof machinevisiontasksislargeanddifficulttotransmitefficiently,a multi-scale imagefeaturefusioncompressionalgorithm wasproposed.Firstly acascadedresidualtransformationmodule wasdesignedaccordingtothe multi-scalefeaturesoutputbythedeeplearning modelontheedgedevice theredundancyofmulti-scalefeatureswaseliminatedbystepwisesubtractionoffeaturesofdifferentsizes andtheresidualfeatureswerecompressedtoaunifiedsize. Then anautoencoderwasdesignedtoeliminatethestatisticalredundancyofcompactfeaturesby arithmeticcoding.Next a prediction andreconstruction module was designed onthecloud accordingtothecompactfeaturesofdecodingtogeneratethepredictionfeatures which were combinedwiththeresidualfeaturestoaccuratelyreconstructthemulti-scalefeatures.Finally,a jointoptimizationfunction wasbuiltforthecollaborativeoptimizationofthe modulesincluding residualtransformation autoencoder andpredictionreconstruction thusachievingtheoptimal trade-offbetweentransmissionbitrateandinformationrepresentationability.Thesimulation resultsshowthattheproposedalgorithmhasnotonlythelargestfeaturecompressionratio but alsothemostcompletereconstructedfeaturesinthespacecompression andthatwhenthetransmissionbitrateis0.1bpp themodelaccuracyoftheproposedalgorithmisimprovedby8.57% and3.87% respectively comparedwiththeimagecodingalgorithm VVCandthefeaturecompressionalgorithm MSFC.Thisstudycanprovidetechnicalsupportforthecodingframeworkof machinevision andhascertainvalueinengineeringapplication.
| Translated title of the contribution | Image Multi-Scale Feature Compression Algorithm for Machine Vision Tasks |
|---|---|
| Original language | Chinese (Traditional) |
| Pages (from-to) | 1-10 |
| Number of pages | 10 |
| Journal | Hsi-An Chiao Tung Ta Hsueh/Journal of Xi'an Jiaotong University |
| Volume | 57 |
| Issue number | 12 |
| DOIs | |
| State | Published - Dec 2023 |