内容中心（MDX）

医学影像分割中 Transformer 与 CNN 融合方法研究

围绕医学分割展开，讨论方法设计、评估指标与工程可用性。

生成对抗网络在人脸修复中的稳定训练策略

围绕GAN 修复展开，讨论方法设计、评估指标与工程可用性。

遥感图像变化检测的时空特征建模方法

围绕遥感变化检测展开，讨论方法设计、评估指标与工程可用性。

结合频域先验的单图像去噪方法研究

围绕频域去噪展开，讨论方法设计、评估指标与工程可用性。

开集识别场景下的图像分类鲁棒性研究

围绕开集识别展开，讨论方法设计、评估指标与工程可用性。

面向移动端的实时图像分割网络架构设计

围绕移动端分割展开，讨论方法设计、评估指标与工程可用性。

弱监督语义分割中伪标签质量提升策略研究

围绕弱监督分割展开，讨论方法设计、评估指标与工程可用性。

细粒度图像识别中的部件建模与关系推理方法

围绕细粒度识别展开，讨论方法设计、评估指标与工程可用性。

三维重建中的神经辐射场方法及改进方向

围绕NeRF 重建展开，讨论方法设计、评估指标与工程可用性。

视觉问答任务中的多模态融合与推理方法

围绕视觉问答展开，讨论方法设计、评估指标与工程可用性。

多尺度注意力机制在图像超分辨率中的应用

围绕图像超分展开，讨论方法设计、评估指标与工程可用性。

面向视频监控的多目标跟踪算法研究与改进

围绕多目标跟踪展开，讨论方法设计、评估指标与工程可用性。

面向低光场景的图像增强网络设计与性能评估

围绕低光增强展开，讨论方法设计、评估指标与工程可用性。

图像检索中的局部特征聚合与重排序方法研究

围绕图像检索展开，讨论方法设计、评估指标与工程可用性。

轻量化目标检测模型在边缘设备上的部署优化

围绕边缘检测展开，讨论方法设计、评估指标与工程可用性。

基于知识蒸馏的图像分类模型压缩方法研究

围绕模型压缩展开，讨论方法设计、评估指标与工程可用性。

面向复杂天气的图像去雾与去雨联合学习框架

围绕复杂天气恢复展开，讨论方法设计、评估指标与工程可用性。

图像描述生成中视觉注意力机制的可解释性分析

围绕图像描述展开，讨论方法设计、评估指标与工程可用性。

数据增广策略对图像分类泛化能力的影响研究

围绕数据增广展开，讨论方法设计、评估指标与工程可用性。

基于扩散先验的图像修复与缺失区域补全研究

围绕图像补全展开，讨论方法设计、评估指标与工程可用性。

基于扩散模型的高保真图像生成与可控编辑研究

围绕扩散模型展开，讨论方法设计、评估指标与工程可用性。

面向工业缺陷检测的小样本学习模型研究

围绕工业质检展开，讨论方法设计、评估指标与工程可用性。

面向隐私保护的联邦学习图像识别方法研究

围绕联邦学习视觉展开，讨论方法设计、评估指标与工程可用性。

自监督学习在遥感图像分类中的有效性分析

围绕自监督遥感展开，讨论方法设计、评估指标与工程可用性。

基于对比学习的图像表征学习方法与应用

围绕对比学习展开，讨论方法设计、评估指标与工程可用性。

图像风格迁移中的内容保持与风格一致性优化

围绕风格迁移展开，讨论方法设计、评估指标与工程可用性。

多任务学习在图像分割与深度估计中的协同优化

围绕多任务学习展开，讨论方法设计、评估指标与工程可用性。

视觉显著性检测中的边界细化方法研究

围绕显著性检测展开，讨论方法设计、评估指标与工程可用性。

对抗样本攻击与图像模型防御方法综述

围绕对抗鲁棒性展开，讨论方法设计、评估指标与工程可用性。

Transformer-CNN Fusion for Medical Image Segmentation

A study on medical segmentation, covering method design, evaluation metrics, and practical usability.

Stable Training Strategies for GAN-based Face Restoration

A study on GAN face restoration, covering method design, evaluation metrics, and practical usability.

Spatiotemporal Modeling for Remote Sensing Change Detection

A study on remote sensing change detection, covering method design, evaluation metrics, and practical usability.

Single-Image Denoising with Frequency-Domain Priors

A study on frequency-aware denoising, covering method design, evaluation metrics, and practical usability.

Robust Image Classification for Open-Set Recognition

A study on open-set recognition, covering method design, evaluation metrics, and practical usability.

Real-Time Image Segmentation Architectures for Mobile

A study on mobile segmentation, covering method design, evaluation metrics, and practical usability.

Pseudo-Label Quality Improvement in Weakly Supervised Segmentation

A study on weakly supervised segmentation, covering method design, evaluation metrics, and practical usability.

Part Modeling and Relational Reasoning for Fine-Grained Recognition

A study on fine-grained recognition, covering method design, evaluation metrics, and practical usability.

Neural Radiance Fields for 3D Reconstruction and Improvements

A study on NeRF reconstruction, covering method design, evaluation metrics, and practical usability.

Multimodal Fusion and Reasoning in Visual Question Answering

A study on visual question answering, covering method design, evaluation metrics, and practical usability.

Multi-Scale Attention for Image Super-Resolution

A study on super-resolution, covering method design, evaluation metrics, and practical usability.

Multi-Object Tracking for Video Surveillance

A study on multi-object tracking, covering method design, evaluation metrics, and practical usability.

Low-Light Image Enhancement Network Design and Evaluation

A study on low-light enhancement, covering method design, evaluation metrics, and practical usability.

Local Feature Aggregation and Re-ranking for Image Retrieval

A study on image retrieval, covering method design, evaluation metrics, and practical usability.

Lightweight Object Detection Optimization on Edge Devices

A study on edge detection deployment, covering method design, evaluation metrics, and practical usability.

Knowledge Distillation for Image Classification Compression

A study on model compression, covering method design, evaluation metrics, and practical usability.

Joint Learning Framework for Dehazing and Deraining

A study on adverse-weather restoration, covering method design, evaluation metrics, and practical usability.

Interpretability of Visual Attention in Image Captioning

A study on image captioning, covering method design, evaluation metrics, and practical usability.

Impact of Data Augmentation on Classification Generalization

A study on data augmentation, covering method design, evaluation metrics, and practical usability.

Image Inpainting with Diffusion Priors

A study on image inpainting, covering method design, evaluation metrics, and practical usability.

High-Fidelity Image Generation and Controllable Editing with Diffusion Models

A study on diffusion models, covering method design, evaluation metrics, and practical usability.

Few-Shot Learning for Industrial Defect Detection

A study on industrial defect detection, covering method design, evaluation metrics, and practical usability.

Federated Learning for Privacy-Preserving Image Recognition

A study on federated vision learning, covering method design, evaluation metrics, and practical usability.

Effectiveness of Self-Supervised Learning in Remote Sensing Classification

A study on self-supervised remote sensing, covering method design, evaluation metrics, and practical usability.

Contrastive Learning Methods for Visual Representation

A study on contrastive representation learning, covering method design, evaluation metrics, and practical usability.

Content Preservation and Style Consistency in Style Transfer

A study on style transfer, covering method design, evaluation metrics, and practical usability.

Co-Optimization of Segmentation and Depth via Multi-Task Learning

A study on multi-task learning, covering method design, evaluation metrics, and practical usability.

Boundary Refinement in Visual Saliency Detection

A study on saliency detection, covering method design, evaluation metrics, and practical usability.

Adversarial Attacks and Defenses in Vision Models

A study on adversarial robustness, covering method design, evaluation metrics, and practical usability.