Detailed Technical Program

Oral Session 1: Image Analysis and Understanding

Talk 1 | 11:00 – 11:18
29: Logit Mixture Outlier Exposure for Fine-grained Out-of-Distribution Detection

Talk 2 | 11:18 – 11:36
57: 2D–3D Feature Fusion via Cross-Modal Latent Synthesis and Attention-Guided Restoration for Industrial Anomaly Detection

Talk 3 | 11:36 – 11:54
61: Complex Facial Expression Recognition Using Deep Knowledge Distillation of Basic Features

Talk 4 | 11:54 – 12:12
71: Evaluating Forgetting in Pretrained Robotic Policy Networks: A Continual Learning Study with Octo

Talk 5 | 12:12 – 12:30
85: Can Current AI Models Count What We Mean, Not What They See? A Benchmark and Systematic Evaluation

Oral Session 2: Semantic Segmentation and Generative Model

Talk 1 | 14:00 – 14:18
2: AWF: Adaptive Weight Fusion for Enhanced Class Incremental Semantic Segmentation

Talk 2 | 14:18 – 14:36
9: Filling the Gaps: A Multitask Hybrid Multiscale Generative Framework for Missing Modality in Remote Sensing Semantic Segmentation

Talk 3 | 14:36 – 14:54
21: Diffusion-Based Synthetic Brightfield Microscopy Images for Enhanced Single Cell Detection

Talk 4 | 14:54 – 15:12
24: Streaming Video Diffusion: Online Video Editing with Diffusion Models

Talk 5 | 15:12 – 15:30
43: Flow-Selectivity SSM: A Generative State-Space Model with History-Aware GFlowNet Policiesn

Oral Session 3: 3D Vision

Talk 1 | 16:00 – 16:18
25: Dynamic Adaptive Sampling for Accurate Image-based 3D Insect Reconstruction Using Neural Implicit Surfaces

Talk 2 | 16:18 – 16:36
32: Consistent3D: Diffusion-Driven Sparse View Completion and 3D Reconstruction with Geometric Priors

Talk 3 | 16:36 – 16:54
62: Beyond Random Masking: A Dual-Stream Approach for Rotation-Invariant Point Cloud Masked Autoencoders

Talk 4 | 16:54 – 17:12
87: Training-free Non-Rigid Registration of Articulated Animal Bodies via Vision Features and Anatomical Priors

Talk 5 | 17:12 – 17:30
100: IntentFuse: Language-Guided 3D Scene Understanding via Prompt Filtering and Fusion

Oral Session 4: Medical Image Analysis

Talk 1 | 11:00 – 11:18
20: From Healthy Scans to Annotated Tumors: A Tumor Fabrication Framework for 3D Brain MRI Synthesis

Talk 2 | 11:18 – 11:36
26: Optimizing Brain Tumor Segmentation Networks through Evolutionary Neural Architecture Search

Talk 3 | 11:36 – 11:54
41: Implicit Visual Chain-of-Thought Reasoning for Anatomically Consistent Unsupervised 4D CT Deformable Image Registration

Talk 4 | 11:54 – 12:12
47: Understanding Non-verbal Vocalizations from Minimally-verbal Autistic Individuals: A Transfer Learning Approach

Talk 5 | 12:12 – 12:30
54: MedAttnNet: A Novel Architecture for Precise Brain Cancer Classification

Oral Session 5: Biomedical

Talk 1 | 11:15 – 11:33
11: Physics-Informed Operator Learning for Hemodynamic Modeling

Talk 2 | 11:33 – 11:51
15: Exploring Primitive Visual Measurement Understanding and the Role of Output Format in Learning in Vision-Language Models

Talk 3 | 11:51 – 12:09
63: mmDFC: Multi-modality Dynamic Fusion and Modal-decoupled Alternating Optimization for Retinal Vessel Occlusion Prognosis

Talk 4 | 12:09 – 12:27
95: Filling in the Missing Piece: Advancing Automated Radiology Report Generation with Clinical Insights

Talk 5 | 12:27 – 12:45
97: X-Gen: Enhancing Radiology Report Generation via LLM-Driven Data Augmentation and Decoupled Training

Oral Session 6: Additional Topics

Talk 1 | 14:00 – 14:18
30: MoE-Enhanced-TTT: Advancing Medical Image Segmentation

Talk 2 | 14:18 – 14:36
53: Maritime Vessel Classification in Low-resolution Synthetic Aperture Radar Imagery

Talk 3 | 14:36 – 14:54
55: Weakly Supervised Pixel-wise Classification of Hyperspectral Images with Noise-Adaptive Hybrid Attention and Triple Contrastive Learning

Talk 4 | 14:54 – 15:12
83: Split-Fuse-Transport: Annotation-Free Saliency via Dual Clustering and Optimal Transport Alignment

Talk 5 | 15:12 – 15:30
104: MedGemma-Critic: Fine-tuning Medical Language Models for Domain-specialised Text Evaluation