Understanding bounding box regression, YOLO (You Only Look Once), SSD (Single Shot MultiBox Detector)

Using OpenCV for detection, Loading pre-trained models (YOLOv8, Faster R-CNN), Running inference on custom images

Artistic style transfer (e.g., Van Gogh style), Content and style separation (Gatys et al.)

Implementing neural style transfer, Optimizing for style-content balance

Photo editing apps, Social media filters, Real-time video stylization

Dimensionality reduction with AEs, Generating images with VAEs

Generating synthetic images with GANs, StyleGAN for high-res generation, CycleGAN for unpaired image translation

Pix2Pix for paired image translation, CycleGAN for style transfer

SimCLR framework, MoCo (Momentum Contrast)

MAE (Masked Autoencoders), BEiT (BERT for Images)

Improving Image Recognition, Reducing labeled data dependency