GitHub - Syun1208/selected-topics-nycu: This course will provide knowledges about computer vision and pattern recognition, including image classification, object detection, instance segmentation, image restoration. These foundations are built by NYCU Selected Topics in Visual Recognition course.

Course: NYCU Selected Topics in Visual Recognition (CV2026)

Assignment	Topics	Best Performance	Checklist
HW1	Image Classification - 100-class ImageNet-derived, ResNet-family (≤100M params)	97.00% val accuracy	✅
HW2	Digit Detection - DETR-family (RT-DETRv2 / DINO / AlignDETR)	51.23 mAP	✅
HW3	Instance Segmentation — Cascade Mask R-CNN / HTC (cell microscopy)	AP50 0.6163 (rank #5)	✅
HW4	Image Restoration - PromptIR, all-in-one rain + snow removal	32.95 dB PSNR	✅
Final Project	Image Matching Challenge 2025 - MASt3R-Hybrid SfM pipeline	56.00 Kaggle (ties 1st place)	✅

Name		Name	Last commit message	Last commit date
Latest commit History 8 Commits
final-project		final-project
hw1		hw1
hw2		hw2
hw3		hw3
hw4		hw4
.gitignore		.gitignore
README.md		README.md