SlideShare a Scribd company logo

Challenges of Deep Learning in Computer Vision Webinar - Tessellate Imaging

•
•
A
Adhesh Shrivastava

Slides from the webinar on Challenges of Deep Learning in Computer Vision presented by Tessellate Imaging and powered by E2E Networks. The webinar discusses the growth and applications of Computer Vision in modern-day real life. Challenges with implementing and developing Deep Learning and Computer Vision projects for both enterprises and developers. We introduce MonkAI (https://monkai.org) an Open Sourced Deep Learning wrapper library for Computer Vision development and talk about features tackling some of the challenges in Deep Learning.

Related slideshows

Real Time Moving Object Detection for Day-Night Surveillance using AI
Real Time Moving Object Detection for Day-Night Surveillance using AIReal Time Moving Object Detection for Day-Night Surveillance using AI
Real Time Moving Object Detection for Day-Night Surveillance using AI
 
Enhance your java applications with deep learning using deep netts
Enhance your java applications with deep learning using deep nettsEnhance your java applications with deep learning using deep netts
Enhance your java applications with deep learning using deep netts
 
OpenPOWER/POWER9 AI webinar
OpenPOWER/POWER9 AI webinar OpenPOWER/POWER9 AI webinar
OpenPOWER/POWER9 AI webinar
 
1 of 23
Download to read offline
Challenges of Deep Learning in Computer Vision Webinar - Tessellate Imaging
How do we see?Why? Source : https://idyll.pub/post/the-eye-5b169094cce3bece5d95e964/
Early applications of Image processing - noise removal - media compression - medical imaging - manufacturing Source : https://progmohamedali.wordpress.com/2014/02/24/image-filters-noise-removal-in-image-processing/
Challenges of Deep Learning in Computer Vision Webinar - Tessellate Imaging
Challenges of Deep Learning in Computer Vision Webinar - Tessellate Imaging
Source : Gartner Symposium India Build AI Business Case 2019 Source : The Forrester New Wave™ - Computer Vision Platforms, Q4 2019 $49Bn Industry by 2023 Growing at a rate of 32% CAGR
Source : https://hai.stanford.edu/sites/g/files/sbiybj10986/f/ai_index_2019_report.pdf
Challenges with Deep Learning Projects
Challenges of Deep Learning in Computer Vision Webinar - Tessellate Imaging
Background Designed By Computer Vision Professionals And Consultants We have been in the domain of Computer Vision industry for the past 7 years, working with a broad spectrum of imaging modalities. Our teams have served clients across the globe solving computational challenges for Computer Vision products. From our learnings, we bring to you a set of easy to use tools for building Computer Vision applications.
Dataset Algorithm (Train + Infer) Deploy Computer Vision + Deep Learning What does it require Compute Power
Let’s check out some state of the art work in the domain of Computer Vision https://paperswithcode.com/sota
Efficient Det (TF) Cornernet (Pytorch) ● Implemented using separate base frameworks Tensorflow & Pytorch ● Different set of dependencies and setup instructions ● Leads to more time spent in prototyping and experimentation ● Working on different projects is of immense cognitive load
Are paid tools an alternative?
MonkAI Simple. Invariant. Unified. Standardising Computer Vision Workflows Powered by Deep Learning Algorithms
MonkAI ● Standard syntax, unified wrapper. Current Support : Pytorch, Keras, MXNet ● Transfer Learning based custom Image Classification ● SOTA Deep Neural Network based Object Detection workflows ● Custom Neural Network Building and Debugging ● Monk-Studio - GUI based Deep Learning Coming Soon -- One-click deploy to cloud, GPU optimisations, Image segmentations, GANs, support for multiple imaging modalities, paper to code and many more.
Image Classification Image Classification -- Pytorch Demo - Create Projects and Experiments - Prepare Dataset (Using foldered or CSV labelled ground truth) - Select pre-trained Deep Neural Network - Resume experiments from the last epoch - Apply layers, activations, tune hyperparameters. - Compare experiments to select the best algorithm - Infer on single or batch of images Blogs Tutorials
Custom Neural Network Builder
Object Detection - Easy to Set Up - Finetune Deep Neural Networks - Github Repo - Documentation and Tutorials Existing Features and available options
Monk-Studio - Github - Image Classification Demo - Object Detection Demo
Going ahead what’s the plan? -- - https://li8bot.github.io/monkai/#/home/demos - https://li8bot.github.io/monkai/#/home/detection/tutorials - https://github.com/Tessellate-Imaging What next?
How should students go about building skills? Some rare lectures available on Youtube : - Image Processing : EENG 512 - Computer Vision -- Colorado School of Mines, Golden, Colorado - Computer Vision : The ancient secrets of Computer Vision
Feel free to reach out to any of our social media channels Linkedin : tessellate-imaging Twitter : @tessellate_img Github : http://bit.ly/monkai-github Website : https://monkai.org https://www.tessellateimaging.com Thank You!

More Related Content

Challenges of Deep Learning in Computer Vision Webinar - Tessellate Imaging

  • 2. How do we see?Why? Source : https://idyll.pub/post/the-eye-5b169094cce3bece5d95e964/
  • 3. Early applications of Image processing - noise removal - media compression - medical imaging - manufacturing Source : https://progmohamedali.wordpress.com/2014/02/24/image-filters-noise-removal-in-image-processing/
  • 6. Source : Gartner Symposium India Build AI Business Case 2019 Source : The Forrester New Wave™ - Computer Vision Platforms, Q4 2019 $49Bn Industry by 2023 Growing at a rate of 32% CAGR
  • 8. Challenges with Deep Learning Projects
  • 10. Background Designed By Computer Vision Professionals And Consultants We have been in the domain of Computer Vision industry for the past 7 years, working with a broad spectrum of imaging modalities. Our teams have served clients across the globe solving computational challenges for Computer Vision products. From our learnings, we bring to you a set of easy to use tools for building Computer Vision applications.
  • 11. Dataset Algorithm (Train + Infer) Deploy Computer Vision + Deep Learning What does it require Compute Power
  • 12. Let’s check out some state of the art work in the domain of Computer Vision https://paperswithcode.com/sota
  • 13. Efficient Det (TF) Cornernet (Pytorch) ● Implemented using separate base frameworks Tensorflow & Pytorch ● Different set of dependencies and setup instructions ● Leads to more time spent in prototyping and experimentation ● Working on different projects is of immense cognitive load
  • 14. Are paid tools an alternative?
  • 15. MonkAI Simple. Invariant. Unified. Standardising Computer Vision Workflows Powered by Deep Learning Algorithms
  • 16. MonkAI ● Standard syntax, unified wrapper. Current Support : Pytorch, Keras, MXNet ● Transfer Learning based custom Image Classification ● SOTA Deep Neural Network based Object Detection workflows ● Custom Neural Network Building and Debugging ● Monk-Studio - GUI based Deep Learning Coming Soon -- One-click deploy to cloud, GPU optimisations, Image segmentations, GANs, support for multiple imaging modalities, paper to code and many more.
  • 17. Image Classification Image Classification -- Pytorch Demo - Create Projects and Experiments - Prepare Dataset (Using foldered or CSV labelled ground truth) - Select pre-trained Deep Neural Network - Resume experiments from the last epoch - Apply layers, activations, tune hyperparameters. - Compare experiments to select the best algorithm - Infer on single or batch of images Blogs Tutorials
  • 19. Object Detection - Easy to Set Up - Finetune Deep Neural Networks - Github Repo - Documentation and Tutorials Existing Features and available options
  • 20. Monk-Studio - Github - Image Classification Demo - Object Detection Demo
  • 21. Going ahead what’s the plan? -- - https://li8bot.github.io/monkai/#/home/demos - https://li8bot.github.io/monkai/#/home/detection/tutorials - https://github.com/Tessellate-Imaging What next?
  • 22. How should students go about building skills? Some rare lectures available on Youtube : - Image Processing : EENG 512 - Computer Vision -- Colorado School of Mines, Golden, Colorado - Computer Vision : The ancient secrets of Computer Vision
  • 23. Feel free to reach out to any of our social media channels Linkedin : tessellate-imaging Twitter : @tessellate_img Github : http://bit.ly/monkai-github Website : https://monkai.org https://www.tessellateimaging.com Thank You!