https://caiyf03.github.io/tags/deep-learning/Deep LearningHugoBlox Kit (https://hugoblox.com)en-usMon, 22 Jan 2024 00:00:00 +0000https://caiyf03.github.io/media/icon_hu_702a800cd775dbac.pnghttps://caiyf03.github.io/tags/deep-learning/https://caiyf03.github.io/projects/od-cs2/Mon, 22 Jan 2024 00:00:00 +0000https://caiyf03.github.io/projects/od-cs2/<h2 id="team">Team</h2> <ul> <li>YiFan Cai (leader)</li> <li>XiHe Yu</li> <li>Yu Shi</li> </ul> <h2 id="overview">Overview</h2> <p>This project investigates the application of real-time object detection in Counter-Strike 2 (CS2). We build a custom in-game dataset and systematically compare multiple detection frameworks to evaluate their accuracy, speed, and practicality in dynamic gaming scenarios. Beyond benchmarking, we demonstrate how detection results can be integrated into gameplay-related applications such as distance estimation and automated aiming.</p> <h2 id="methods">Methods</h2> <p>We implement and compare YOLOv7, Faster R-CNN, and SSD. Key components include:</p> <p>-Construction of a custom VOC-style CS2 dataset from gameplay footage</p> <p>-Model training and evaluation under identical settings</p> <p>-Preprocessing and data augmentation for difficult scenes</p> <p>-Ensemble-style bounding box fusion using confidence weighting and clustering</p> <h2 id="key-features">Key Features</h2> <p>-Custom-built CS2 object detection dataset</p> <p>-Comparative study of one-stage vs. two-stage detectors</p> <p>-Model ensemble via weighted bounding box averaging</p> <p>-Real-time in-game deployment with screen capture</p> <p>-Applications including auto-aiming and target distance estimation</p> <hr> <figure style="text-align: center; margin-bottom: 2rem;"> <img src="1.png" alt="" style="width: 100%; max-width: 650px;" /> <figcaption style="margin-top: 0.5rem; font-size: 0.9rem; color: #555;"> </figcaption> </figure> <figure style="text-align: center; margin-bottom: 2rem;"> <img src="2.png" alt="" style="width: 100%; max-width: 650px;" /> <figcaption style="margin-top: 0.5rem; font-size: 0.9rem; color: #555;"> </figcaption> </figure>https://caiyf03.github.io/internships/shiye/Thu, 01 Sep 2022 00:00:00 +0000https://caiyf03.github.io/internships/shiye/<h2 id="overview">Overview</h2> <p><strong>Research Member</strong> -ShanghaiTech University, 2022 – 2025*</p> <h3 id="generative-modeling-research">Generative Modeling Research</h3> <ul> <li>Conducted research on diffusion-based generative models for image synthesis, image restoration, and controllable generation.</li> <li>Participated in projects spanning computer vision and small-molecule generation under structured optimization objectives.</li> <li>Explored model architecture design, training strategies, and generation-quality enhancement techniques.</li> </ul> <h3 id="academic-research-workflow">Academic Research Workflow</h3> <ul> <li>Participated in the complete lifecycle of top-tier conference publications, including literature review, problem formulation, model design, experimentation, and manuscript preparation.</li> <li>Co-authored a peer-reviewed publication as the third author.</li> <li>Accumulated three years of continuous research experience in generative modeling and deep learning.</li> </ul> <h3 id="experimental-development">Experimental Development</h3> <ul> <li>Reproduced and extended recent research papers in diffusion models and generative learning.</li> <li>Designed model improvements and conducted hyperparameter optimization for training stability and generation quality.</li> <li>Performed systematic analysis of model convergence behavior and output fidelity.</li> </ul> <hr>