{"paper":{"title":"Towards Label-Free Single-Cell Phenotyping Using Multi-Task Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A deep learning model jointly classifies white blood cell types and regresses protein expression levels from label-free DPC images.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Ardhendu Behera, Saqib Nazir","submitted_at":"2026-05-14T11:38:56Z","abstract_excerpt":"Label-free single-cell imaging offers a scalable, non-invasive alternative to fluorescence-based cytometry, yet inferring molecular phenotypes directly from bright-field morphology remains challenging. We present a unified Deep Learning (DL) framework that jointly performs White Blood Cell (WBC) classification and continuous protein-expression regression from label-free Differential Phase Contrast (DPC) images. Our model employs a Hybrid architecture that fuses convolutional fine-grained texture features with transformer-based global representations through a learnable cross-branch gating modu"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We present a unified Deep Learning (DL) framework that jointly performs White Blood Cell (WBC) classification and continuous protein-expression regression from label-free Differential Phase Contrast (DPC) images.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that bright-field morphology in DPC images contains enough information to accurately infer molecular phenotypes such as protein expression levels.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A hybrid CNN-transformer model with multi-task learning achieves 91.3% WBC classification accuracy and 0.72 Pearson correlation for CD16 expression regression from label-free DPC images, augmented by LLM-generated summaries.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A deep learning model jointly classifies white blood cell types and regresses protein expression levels from label-free DPC images.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"8672b1b6cdcfe1f4fa9cc060f303d745bcb8911838db0ceb495395b43ade0f41"},"source":{"id":"2605.14717","kind":"arxiv","version":1},"verdict":{"id":"2da290a6-b355-4686-8987-fcc474ab480f","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T04:39:01.992450Z","strongest_claim":"We present a unified Deep Learning (DL) framework that jointly performs White Blood Cell (WBC) classification and continuous protein-expression regression from label-free Differential Phase Contrast (DPC) images.","one_line_summary":"A hybrid CNN-transformer model with multi-task learning achieves 91.3% WBC classification accuracy and 0.72 Pearson correlation for CD16 expression regression from label-free DPC images, augmented by LLM-generated summaries.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that bright-field morphology in DPC images contains enough information to accurately infer molecular phenotypes such as protein expression levels.","pith_extraction_headline":"A deep learning model jointly classifies white blood cell types and regresses protein expression levels from label-free DPC images."},"references":{"count":28,"sample":[{"doi":"","year":2021,"title":"TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation","work_id":"ffa2ac60-2755-4390-9f66-07815aa6cb27","ref_index":1,"cited_arxiv_id":"2102.04306","is_internal_anchor":true},{"doi":"","year":2024,"title":"Lab on a Chip24(5), 924–932 (2024)","work_id":"6b031fbd-eaeb-4508-9136-6c23cefb7746","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2016,"title":"arXiv: Learning (2016),https: //api.semanticscholar.org/CorpusID:125617073","work_id":"bec233d0-3795-418c-8835-b4049859a591","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"In: International Conference on Learning Representations (2021)","work_id":"698c0358-c2a8-4374-be3b-30929069d2a6","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Google DeepMind: Gemini 2.5 pro model card.https://deepmind.google/(2024) Towards Label-Free Single-Cell Phenotyping Using Multi-Task Learning 15","work_id":"ba3a424d-fd17-4f6b-b501-b6e2bb83d7a9","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":28,"snapshot_sha256":"6299e79ae1b5fdee89d4f450aeae3b462a6dfdc7bc350a0583fc5330018576e9","internal_anchors":2},"formal_canon":{"evidence_count":2,"snapshot_sha256":"fcb45373833722716f9e01dc02841b62e3f9de2b60554f92da8c470070ab63d1"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}