paper-with-me

홈 › Papers

LoMix: Learnable Weighted Multi-Scale Logits Mixing for Medical Image Segmentation

2025-10-27 · Md Mostafijur Rahman, Radu Marculescu arxiv

U-shaped networks output logits at multiple spatial scales, each capturing a different blend of coarse context and fine detail. Yet, training still treats these logits in isolation - either supervising only the final, highest-resolution logits or applying deep supervision with identical loss weights at every scale - without exploring mixed-scale combinations. Consequently, the decoder output misses the complementary cues that arise only when coarse and fine predictions are fused. To address this issue, we introduce LoMix (Logits Mixing), a NAS-inspired, differentiable plug-and-play module that generates new mixed-scale outputs and learns how exactly each of them should guide the training process. More precisely, LoMix mixes the multi-scale decoder logits with four lightweight fusion operators: addition, multiplication, concatenation, and attention-based weighted fusion, yielding a rich set of synthetic mutant maps. Every original or mutant map is given a softplus loss weight that is co-optimized with network parameters, mimicking a one-step architecture search that automatically discovers the most useful scales, mixtures, and operators. Plugging LoMix into recent U-shaped architectures (i.e., PVT-V2-B2 backbone with EMCAD decoder) on Synapse 8-organ dataset improves DICE by +4.2% over single-output supervision, +2.2% over deep supervision, and +1.5% over equally weighted additive fusion, all with zero inference overhead. When training data are scarce (e.g., one or two labeled scans), the advantage grows to +9.23%, underscoring LoMix's data efficiency. Across four benchmarks and diverse U-shaped networks, LoMiX improves DICE by up to +13.5% over single-output supervision, confirming that learnable weighted mixed-scale fusion generalizes broadly while remaining data efficient, fully interpretable, and overhead-free at inference. Our code is available at https://github.com/SLDGroup/LoMix.

📄 PDF Abstract BibTeX arXiv:2510.22995

Code (0)

등록된 구현이 없습니다.

Tasks

Medical Image Segmentation

Similar Papers 제목 키워드 기반

oculomix: Hierarchical Sampling for Retinal-Based Systemic Disease Prediction

2026-01-16 · Hyunmin Kim, Yukun Zhou, Rahul A. Jonas, Lie Ju 외 arxiv

Oculomics - the concept of predicting systemic diseases, such as cardiovascular disease and dementia, through retinal imaging - has advanced rapidly due to the data efficiency of transformer-based foundation models like …

An MIL-Derived Transformer for Weakly Supervised Point Cloud Segmentation

2022-01-01 · CVPR 2022 1 · Cheng-Kun Yang, Ji-Jia Wu, Kai-Syun Chen, Yung-Yu Chuang 외

We address weakly supervised point cloud segmentation by proposing a new model, MIL-derived transformer, to mine additional supervisory signals. First, the transformer model is derived based on multiple instance lear…

Model OptimizationMultiple Instance LearningPoint Cloud Segmentation

Evaluating the Robustness of Geometry-Aware Instance-Reweighted Adversarial Training

2021-03-02 · Dorjan Hitaj, Giulio Pagnotta, Iacopo Masi, Luigi V. Mancini

In this technical report, we evaluate the adversarial robustness of a very recent method called "Geometry-aware Instance-reweighted Adversarial Training"[7]. GAIRAT reports state-of-the-art results on defenses to adversa…

Adversarial AttackAdversarial Robustness

High-Fidelity Text-to-Image Generation from Pre-Trained Vision-Language Models via Distribution-Conditioned Diffusion Decoding

2026-03-11 · Ji Woo Hong, Hee Suk Yoon, Gwanhyeong Koo, Eunseop Yoon 외 arxiv

Recent large-scale vision-language models (VLMs) have shown remarkable text-to-image generation capabilities, yet their visual fidelity remains constrained by the discrete image tokenization, which poses a major challeng…

Text-to-Image Generation

Mitigating Measurement-Induced Training Instability in Hybrid Quantum Neural Networks for Protein Classification

2026-06-21 · Milton Mondal, Sushovan Chanda, Mohamad Mahdi Alawieh, Brijesh Sukhadiya 외 arxiv

Hybrid Quantum Neural Network (QNN) classifiers produce logits as expectation values of quantum measurement operators. For standard Pauli measurements, these outputs are intrinsically bounded to the interval [-1,1]. When…

Multi-class Classification