Papers › Trapped in texture bias? A large scale comparison of deep instance segmentation

Trapped in texture bias? A large scale comparison of deep instance segmentation

17 Jan 2024arXiv:2401.09109archive 2025-07-28

Johannes Theodoridis, Jessica Hofmann, Johannes Maucher, Andreas Schilling

Do deep learning models for instance segmentation generalize to novel objects in a systematic way? For classification, such behavior has been questioned. In this study, we aim to understand if certain design decisions such as framework, architecture or pre-training contribute to the semantic understanding of instance segmentation. To answer this question, we consider a special case of robustness and compare pre-trained models on a challenging benchmark for object-centric, out-of-distribution texture. We do not introduce another method in this work. Instead, we take a step back and evaluate a broad range of existing literature. This includes Cascade and Mask R-CNN, Swin Transformer, BMask, YOLACT(++), DETR, BCNet, SOTR and SOLOv2. We find that YOLACT++, SOTR and SOLOv2 are significantly more robust to out-of-distribution texture than other frameworks. In addition, we show that deeper and dynamic architectures improve robustness whereas training schedules, data augmentation and pre-training have only a minor impact. In summary we evaluate 68 models on 61 versions of MS COCO for a total of 4148 evaluations.

PaperPDFCode

Code

johannestheo/trapped-in-texture-bias officialmentioned in paperpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Data AugmentationInstance SegmentationObject DetectionOut-of-Distribution GeneralizationSemantic Segmentation

Datasets

Introduced by this paper, per the archive.

Object-Centric Stylized COCO

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Absolute Position EncodingsAdamAttentionBPEConvolutionDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMask R-CNNMulti-Head AttentionPosition-Wise Feed-Forward LayerRPNResidual ConnectionRoIAlignSoftmaxStochastic DepthSwin TransformerTransformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections