MARIS: Marine Open-Vocabulary Instance Segmentation with Geometric Enhancement and Semantic Alignment

2025/10/17 by Bingyu Li, Feiyu Wang, Li, Bingyu +9 · 2 citations
Computer Science · #Advanced Neural Network Applications #Artificial Intelligence (cs.AI) #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Image Enhancement Techniques #Multimodal Machine Learning Applications

paper · pdf · doi:10.48550/arxiv.2510.15398

openalex publication_date 2025/10/17 · openalex created_date 2025/10/21 · openalex updated_date 2026/07/28

Abstract

Most existing underwater instance segmentation approaches are constrained by close-vocabulary prediction, limiting their ability to recognize novel marine categories. To support evaluation, we introduce MARIS (\underlineMarine Open-Vocabulary \underlineInstance \underlineSegmentation), the first large-scale fine-grained benchmark for underwater Open-Vocabulary (OV) segmentation, featuring a limited set of seen categories and diverse unseen categories. Although OV segmentation has shown promise on natural images, our analysis reveals that transfer to underwater scenes suffers from severe visual degradation (e.g., color attenuation) and semantic misalignment caused by lack underwater class definitions. To address these issues, we propose a unified framework with two complementary components. The Geometric Prior Enhancement Module (GPEM) leverages stable part-level and structural cues to maintain object consistency under degraded visual conditions. The Semantic Alignment Injection Mechanism (SAIM) enriches language embeddings with domain-specific priors, mitigating semantic ambiguity and improving recognition of unseen categories. Experiments show that our framework consistently outperforms existing OV baselines both In-Domain and Cross-Domain setting on MARIS, establishing a strong foundation for future underwater perception research.

Citations

• STARE-VLA: Progressive Stage-Aware Reinforcement for Fine-Tuning Vision-Language-Action Models
• UVLM: Benchmarking Video Language Model for Underwater World Understanding
• Fine-Grained Preference Optimization Improves Spatial Reasoning in VLMs
• USIS16K: High-Quality Dataset for Underwater Salient Instance Segmentation
• Cross from Left to Right Brain: Adaptive Text Dreamer for Vision-and-Language Navigation
• AUTV: Creating Underwater Video Datasets with Pixel-wise Annotations
• Training Data Provenance Verification: Did Your Model Use Synthetic Data from My Generative Model for Training?
• Efficient Redundancy Reduction for Open-Vocabulary Semantic Segmentation
• FGAseg: Fine-Grained Pixel-Text Alignment for Open-Vocabulary Semantic Segmentation
• Unsupervised Hyperspectral and Multispectral Image Fusion via Self-Supervised Modality Decoupling
• Generalization Boosted Adapter for Open-Vocabulary Segmentation
• Collaborative Vision-Text Representation Optimizing for Open-Vocabulary Segmentation
• Open-Vocabulary Semantic Segmentation with Image Embedding Balancing
• Depth Anything V2
• Diving into Underwater: Segment Anything Model Guided Underwater Salient Instance Segmentation and A Large-scale Dataset
• U3M: Unbiased Multiscale Modal Fusion Model for Multimodal Semantic Segmentation
• Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data
• Open-Vocabulary Segmentation with Semantic-Assisted Calibration
• SED: A Simple Encoder-Decoder for Open-Vocabulary Semantic Segmentation
• CaveSeg: Deep Semantic Segmentation and Scene Parsing for Autonomous Underwater Cave Exploration
• Convolutions Die Hard: Open-Vocabulary Segmentation with Single Frozen Convolutional CLIP
• InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning
• FreeSeg: Unified, Universal and Open-Vocabulary Image Segmentation
• CAT-Seg: Cost Aggregation for Open-Vocabulary Semantic Segmentation
• Open-Vocabulary Panoptic Segmentation with Text-to-Image Diffusion Models
• Side Adapter Network for Open-Vocabulary Semantic Segmentation
• BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models
• Open-Vocabulary Semantic Segmentation with Mask-adapted CLIP
• MaskCLIP: Masked Self-Distillation Advances Contrastive Language-Image Pretraining
• BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation
• Language-driven Semantic Segmentation
• A Simple Baseline for Open-Vocabulary Semantic Segmentation with Pre-trained Vision-language Model
• An Underwater Image Semantic Segmentation Method Focusing on Boundaries and a Real Underwater Scene Semantic Segmentation Dataset
• Per-Pixel Classification is Not All You Need for Semantic Segmentation
• Learning Transferable Visual Models From Natural Language Supervision
• Semantic Segmentation of Underwater Imagery: Dataset and Benchmark
• Microsoft COCO: Common Objects in Context

Cited by

• Exploring the Underwater World Segmentation without Extra Training
• The Devil Is in Gradient Entanglement: Energy-Aware Gradient Coordinator for Robust Generalized Category Discovery

Related

• Exploring the Underwater World Segmentation without Extra Training
• FSSUWNet: Mitigating the Fragility of Pre-trained Models with Feature Enhancement for Few-Shot Semantic Segmentation in Underwater Images
• Advancing Marine Research: UWSAM Framework and UIIS10K Dataset for Precise Underwater Instance Segmentation
• Semantic Segmentation of Underwater Imagery: Dataset and Benchmark
• MarineDet: Towards Open-Marine Object Detection