A code-first course on building vision systems with foundation and generative models. Part of the Hands-On AI Science series, designed around Innovation-First Learning principles.
Companion Online Textbook
Vision AI has moved from narrow classifiers to foundation models that understand, describe, and generate images. Every product that touches cameras, documents, or visual content now depends on these capabilities. This course prepares students to build vision systems that see, reason, and create.
Core concepts, models, and ideas behind modern computer vision: convolutional and vision transformer architectures, contrastive learning, diffusion processes, latent space geometry, and multimodal alignment between images and text.
PyTorch, OpenCV, Hugging Face Diffusers, Stable Diffusion, YOLO, SAM, CLIP, Roboflow, Weights & Biases, and Google Colab.
A specific course syllabus is built for each audience: graduate or undergraduate, across engineering, digital health, or computer science.
As AI and mature libraries handle standard tasks, professional developers must focus on innovation. Student projects tackle new use cases by generating unique data and fine-tuning task-specific vision models.
Students begin their projects while learning the material and enrich them as new concepts arrive. Each team gives several in-class presentations for discussion and feedback.
Image Processing Fundamentals
OpenCV, NumPy, filtering, edges
CNNs & Image Classification
PyTorch, ResNet, transfer learning
Object Detection
YOLO, Roboflow, bounding boxes
Semantic Segmentation
SAM, U-Net, pixel-level masks
Project Proposal Presentations
Student proposals, peer feedback
Vision Transformers
ViT, DINOv2, Hugging Face
Multimodal Models & CLIP
CLIP, BLIP-2, visual QA
Interim Project Presentations
Progress demos, instructor feedback
Generative Models: GANs
StyleGAN, image-to-image translation
Diffusion Models
Stable Diffusion, HF Diffusers
Image Editing & ControlNet
Inpainting, ControlNet, img2img
Video Understanding & 3D Vision
Optical flow, depth estimation, NeRF
Final Project Presentations
Live demos, peer evaluation
This modular series is taught as concrete courses. Full syllabi for the specific Vision AI offerings:
A 13-week graduate course on generative modeling, spanning variational autoencoders, diffusion models, and world models.
A 13-week graduate course on deep generative models for audio-visual data, covering VAEs, GANs, diffusion and the Stable Diffusion family, and synthetic data generation.
A 13-week undergraduate course on deep generative models for visual data, covering VAEs, GANs, diffusion and the Stable Diffusion family, and synthetic data generation.
A 13-week undergraduate course tracing computer vision from image formation and classical image processing through classical vision and 3D reconstruction to deep learning for detection and segmentation, taught with a code-first approach.
