Deep Learning – Computer Vision #
Foundation of Computer Vision #
- Common Architectural Principles of Deep Networks
- Building Blocks of Deep Networks
- Convolutional Neural Networks (CNNs)
- Recurrent Neural Networks
- Recursive Neural Networks; Applications to Sequence Data
- Anomaly Detection
- Tuning Deep Networks
- Vectorization
- Data Mining (Pre-requisites)
CNN overview #
- CNN Definition
- CNN based Architectures
- End to end CNN network
- Training CNN
- Deployment in Azure Cloud
- Performance tuning of CNN network
Advance Computer Vision – Part 1 #
- CNN Architectures with research paper and mathematics
- Resnet-5 variants with research paper and practical
- AlexNet variants with research paper and practical
- GoogleNet variants with research paper and practical
- Transfer learning
- VGGNet variants with research paper and practical
- Inception net variants with research paper and practical
- Darknet variants with research paper and practical
Advance Computer Vision – Part 2 #
- Object detection in-depth
- Transfer learning
- RCNN with research paper and practical
- Fast RCNN with research paper and practical
- Faster RCNN with research paper and practical
- SSD with research paper and practical
- SSD lite with research paper and practical
Training of Custom Object Detection #
- TFOD introduction
- Environment setup wtih TFOD
- GPU vs TPU vs CPU
- GPU Comparison
Advance Computer Vision – Part 3 #
- Yolo v1 with research paper and practical
- Retina net
- Face net
- Detectron2 with practical and live testing
Object segmentation #
- Semantic segmentation
- Panoptic segmentation
- Masked RCNN
- Practical with Detectron
- Practical with TFOD
Object tracking #
- Detail of object tracking
- Kalman filtering
- Sort
- Deep sort
- Object tracking live project with live camera testing
OCR #
- Introduction to OCR
- Various framework and API for OCR
- Practical implementation of OCR
- Live project deployment for bill parsing
Image captioning #
- Image captioning overview
- Image captioning project with deployment
Comments: