Deep Learning in Computer Vision
Sayon Dutta (~sayon) |
Description:
In the world of computer vision, the current advancements in recent years have led us to segement images and predict classes of the multiple objects within the image. This talk will cover basic features of OpenCV and traditional image processing steps followed by advancement in the domain of using deep learning. The goal of this talk is to touch base with the current deep learning architectures deployed in the field of computer vision and their application domains.
The outline of the talk covers:
- Key features of OpenCV
- CNN Basics review for Object Classification
- GANs for Image Generation
- Image segmentation and localisation:
- RCNN
- Fast RCNN
- Faster RCNN
- Mask RCNN
- SegNet
- YOLO
- UNet
Prerequisites:
- Basics of linear algebra
- Basics of neural networks
- Intermediate knowledge of Python
- Basics of Convolutional Network
Content URLs:
Work in progress
Speaker Info:
Sayon Dutta has 4 years of work experience in the field of Machine Learning. Major areas of interest being deep learning in NLP and Image Processing. His work experience include:
- Research Scientist @ CropIn (currently)
- AI Engineer @ Wissen
- Co-founder and VP- AI Research @ Marax AI
- Data Scientist @ ZipGo
- Executive Analyst @ Deloitte
He is also the author of the book Reinforcement Learning with Tensorflow and also owns a software copyright for Mobile Irrigation Scheduler
