Search papers, labs, and topics across Lattice.
YOLOv14 introduces a unified framework for real-time object detection that significantly improves performance on challenging non-ideal inputs, such as fisheye images and game-rendered characters. Key innovations include Deformable Area-Attention for adaptive sampling, Game2Real Domain Adaptation for aligning feature distributions, Multi-View Conditioning for enhancing viewpoint representation, and an Adaptive Augmentation Policy for optimal input handling. As a result, YOLOv14 achieves a mean Average Precision (mAP) of 49.1 on COCO val2017, with notable improvements across various challenging benchmarks, including a 26.1 mAP increase for game characters.
YOLOv14 boosts object detection accuracy by over 26 mAP for game characters, transforming how we handle domain adaptation in real-time scenarios.
Real-time object detectors achieve remarkable accuracy under controlled conditions, yet degrade sharply on non-ideal inputs: fisheye distortion, game-rendered characters, aerial viewpoints, and 360{\deg} panoramas. We present YOLOv14, aunified detection framework addressing these challenges through four synergisticinnovations. (1) Deformable Area-Attention (D-AAttn) replaces rigid attentiongrids with learned 2D deformation fields, enabling adaptive sampling under geometric distortion. (2) Game2Real Domain Adaptation aligns rendered-game and photographic feature distributions via Adaptive Instance Normalization (AdaIN)and adversarial domain confusion, allowing game characters are detected as realhumans. (3) Multi-View Conditioning injects learned viewpoint embeddings intothe backbone with a cross-view contrastive loss that pulls same-class features fromdifferent perspectives closer. (4) An Adaptive Augmentation Policy automaticallyclassifies each input'scene type and routes to optimal augmentations, while a DynamicScaleRouter learns per-input feature pyramid weights. Together, YOLOv14achieves 49.1 mAP on COCO val2017 at 2.91 ms (T4 GPU), and delivers substantial gains on fisheye (+4.1 mAP), panorama (+6.6 mAP), drone (+6.4 mAP), andgame-character (+26.1 mAP) benchmarks
