https://www.acal.org.cn/en/tags/image-restoration/Image RestorationHugoBlox Kit (https://hugoblox.com)en-usWed, 01 Jan 2025 00:00:00 +0000https://www.acal.org.cn/media/logo_hu_9c369f2c5735daea.pnghttps://www.acal.org.cn/en/tags/image-restoration/https://www.acal.org.cn/en/research/computer-vision/Wed, 01 Jan 2025 00:00:00 +0000https://www.acal.org.cn/en/research/computer-vision/<h2 id="overview">Overview</h2> <p>We apply deep learning techniques to medical imaging, weather prediction, industrial inspection, and other domains, researching core vision tasks including image segmentation, restoration, generation, and enhancement, with emphasis on combining algorithmic innovation with real-world deployment.</p> <h2 id="research-directions">Research Directions</h2> <ul> <li><strong>Medical Image Segmentation</strong>: Cardiac MRI segmentation and multi-task learning based on deep CNNs</li> <li><strong>Image Restoration &amp; Enhancement</strong>: Multi-scale progressive face deblurring and attribute-transformation-based pluralistic face inpainting</li> <li><strong>Intelligent Weather Prediction</strong>: Near-surface air temperature estimation via improved conditional GANs; reinforcement learning-based meteorological reasoning chain generation</li> <li><strong>Industrial Defect Detection</strong>: Intelligent defect recognition and reconstruction algorithms in electromagnetic non-destructive testing</li> <li><strong>Iris Segmentation</strong>: Precise iris region segmentation based on Mask R-CNN</li> </ul> <h2 id="key-achievements">Key Achievements</h2> <ul> <li>Proposed improved conditional GAN for near-surface air temperature estimation (Sensors 2024)</li> <li>Proposed automated chain-of-thought data generation for meteorological reasoning (ICIMCT 2025)</li> <li>Proposed multi-scale progressive blind face deblurring algorithm (Complex &amp; Intelligent Systems 2022)</li> <li>Proposed pluralistic face inpainting with attribute transformation (IEEE TMM 2022)</li> <li>Proposed hypergraph-enhanced U-Net segmentation method (CSSE 2022)</li> <li>Leading industrial inspection projects on multi-layer electromagnetic-thermal tomography</li> </ul>