Touches of Python in Prosthetic Vision : New Generation Computer vision

Anustup


Description:

Computer Vision developed a lot in the field of Generative learning , Body pose models as well 3D image mapping . Even we can generate inbuild AR/VR by using the power AI and Computer Vision. So this project is also based on similar concept but on the forensic and prosthetic concept .First of all prosthetics are creating similar human masks that create no difference with the actual face not even from the fitting points . Facial prosthetics helps a lot in media as well security industry along with health IT in curing patients with multiple problems including disability ,mental problems as well facial de formations and so on ! Apart from Prosthetics , we are far away developed in the field of argumented and virtual reality! We all are using automated apps as well platforms to create vision into reality .This projects of AI prosthetic vision is developed to integrate the power of computer vision , Face map generator by the light of prosthetics in the power of Face mesh algorithm. This is specially designed to help patients with cognitive impairment and to help police to extract face maps from the face sketch of criminals !

Problem statement :

  1. In the security department police generally have a rough sketch of the criminal , they always want to have a proper vision of the criminal , so If we have a hardware come software that will change the input image into the 3D mated image then that will be great .

  2. The patients who are Divyang that means physically challenged , they loves to see their own body in different colors as well moving and playing in different directions . Though generating a avtar and making this type of activities is possible by Argumented Reality but its difficult to bear it where ever we want to go , so this work will help them to generate this kind of things . So I am creating a camera based system using the power of Computer vision Deep Pose Facebook model that will generate 3D prosthetic model from a 2D image or video . This 3D prosthetic output we can use anywhere to generate different activities as well 3D modelling . Also this 3D modelling will help in virtual face surgery understanding . I will generate a complete hardware and software to create this out of the real time scenario which will help the Divyang , Police security and Face modelling .

So this talk will be speaking about how we can simply add prosthetics to our amazing computer vision algorithms to generate mesh models ,3D architectures etc easily by the power of Python and Python Libraries . Also this talk will throw lights on extensibility of Py ml models with Tensorflow JS and how to scale up that models into production ,optimisation of advanced Facebook AI's models like Detectron 2 by the power of python and python libraries . In short this talk will let audiences know how we can use the power of Python in advanced computer vision called as Generative Learnings ( GAN ) , In built AR VR as well in adding automated prosthetics which can easily solve the million dollar problems stated above.

Prerequisites:

People who will listen this talk , should have to know at least the following to get most of the talk in the functioning of their brain :

  • Basic Understanding of Python and Python Libraries
  • Basic knowledge on how computer vision model's works & how basic AI works.
  • Some knowledge on 3D modelling will be an additional perk !

Video URL:

https://drive.google.com/file/d/1OAPs3JKzOvuajRxsrvzpyP_qPtxfUJNp/view?usp=sharing

Content URLs:

  • Github : https://github.com/Anustup900/AI-Prosthetic-vision
  • Powerpoint : https://www.canva.com/design/DAEdWKJlRX4/P70QXsq_zvfmiuucNg9cQQ/view?utm_content=DAEdWKJlRX4&utm_campaign=designshare&utm_medium=link&utm_source=sharebutton
  • Demo Video of one application : https://drive.google.com/drive/folders/1ohDQZwZ7q3j_thUdZUgvJVuM8z1N_9lA?usp=sharing

Speaker Info:

AI Developer with 2 years of work experience in 5 + companies , writes code to create functional AI & AI for Health IT ! Present computer vision engineer at Dark Horse INC ,USA Silicon valley . Former Lead AI engineer at Omdena USA . Former winter AI intern at IIT Patna , Researched at Makerology labs IIT Bombay . Holding patents ,research & publication in AI ML . Spoken at 20 + national and International conferences including Belpy 21 , GS India Summit 20 , MIT Tech X ,John Hopkins University Med Hacks etc .Trained more than 1K+ students in the field of AI ML and development over an year . Featured in Press and media too for start up MBK Health Tech ,for creating AI based remote cardiac monitor for human health .

Speaker Links:

  • Linkedin : https://www.linkedin.com/in/anustupmukherjee/
  • Github : https://github.com/Anustup900

Section: Data Science, Machine Learning and AI
Type: Talks
Target Audience: Intermediate
Last Updated: