Enhancing Pathology Image Analysis with Semi-Supervised Learning: A ConvNeXt and U-Net Hybrid Framework for Cancer Diagnosis
Badr Elguerch · International Journal for Research in Applied Science and Engineering Technology · 2025
Abstract:Digitalpathologyhasexperiencedsignificantadvancementsduetotheincreasingrelianceoncomputationalmodelsformedical diagnosisanddiseasedetection. However,amajorobstaclepersists: thescarcityofannotateddatasets required for training deep learning models.Manual annotation by medical experts is time-consuming, expensive, and pronetoerrors, limitingthedevelopmentofrobustdiagnosticsystems.Thisresearchaddressesthischallengebyproposingahybridframeworkthatcombin esSemi-SupervisedLearning(SSL)techniqueswithConvNeXtandU-Netarchitectures to enhance cancer diagnosis using limited labeled data.The study employs SSL strategies such as pseudo-labeling and consistency regularizationtomaximizethe useofunlabeleddatawhileimprovingmodelgeneralization. ConvNeXtserves astheencoderforfeatureextraction, whileUNetactsasthedecoderforprecisesegmentationtasks.Dataaugmentationtechniquesfurtherenhancetrainingdiversity,reducingoverfitti ngandimprovinggeneralization.ExperimentalresultsonthePANDAdatasetdemonstratesuperiorperformance,achievingaQuadratic WeightedKappa(QWK)scoreof0.9700,ClinicalAccuracy(ClinAcc)of93%,andAUROCof0.9600.Thesefindingshighlightthepotential ofSSLinovercoming annotation scarcity in digital pathology while paving the way for scalable AI solutions in clinical settings. Keywords: Semi-SupervisedLearning,DigitalPathology,CancerDiagnosis,ConvNeXt-XXL,U-Net