Midv-550 -

Data augmentation (random motion blur, brightness jitter, perspective warp) during OCR training yields a 22 % relative CER reduction. | Pipeline | E2E Accuracy | Composite Score (S) | |----------|--------------|---------------------| | YOLOv8

Existing public benchmarks (e.g., [1], IDDoc [2], SROIE [3]) either contain a limited number of document classes, provide only coarse bounding‑box annotations, or lack realistic mobile acquisition conditions. Consequently, progress in robust MIV systems has been hindered by a mismatch between training data and real‑world deployment scenarios. MIDV-550

A composite score is reported for overall ranking. 5. Experimental Results 5.1 Document Detection | Model | mAP@0.5 | Inference (ms / img) | |-------|---------|----------------------| | Faster R‑CNN (ResNet‑101) | 0.89 | 128 | | EfficientDet‑D4 | 0.92 | 71 | | YOLOv8‑x (baseline) | 0.95 | 38 | A composite score is reported for overall ranking

: Recent works use instance‑segmentation (Mask RCNN [8]) or keypoint‑based approaches (DETR‑Doc [9]) to isolate MRZ, portrait, and signature regions. : Sequence‑to‑sequence models (CRNN [10])

: Sequence‑to‑sequence models (CRNN [10]), Transformer‑based recognizers (SATRN [11]), and large‑scale pre‑trained vision‑language models (TrOCR [12]) have set the state‑of‑the‑art on clean scanned documents but degrade sharply on mobile captures.

YOLOv8‑x attains the highest detection recall (98 %) while maintaining real‑time speed on mobile‑grade CPUs (≈ 150 ms per image using TensorRT). | Model | Mean IoU (all fields) | MRZ IoU | Portrait IoU | |-------|----------------------|----------|--------------| | Mask RCNN (ResNeXt‑101) | 0.78 | 0.84 | 0.71 | | DETR‑Doc (ViT‑B) | 0.74 | 0.80 | 0.68 | | Mask RCNN + Geometric Refine (baseline) | 0.82 | 0.88 | 0.75 |