AI / ML USE CASES
Landfill recycling object identification using Tensor-flow (Python AI/ML solution)
The project “Landfill Refinery” is a Pyhon AI/ML based solution using Google Tensorflow, that can detect object such as, ",Plastic Bottle", “Aluminum Can”, “Banana Peels”, etc. which are recyclable or Bio-Degradable from landfill. Landfill are huge dump areas situated in outskates of the urban areas. This landfills have every type of waste (which is useless for the human) products, that include Bio-Degradable, Toxic, Recyclable, etc. Waste products that are Recyclable and Bio-Degradable can be brought and used for many purposes. At present, manual scavengers do this task. with very high risk (fire and smoke) and also life threatening diseases. s This solution can detect object and classify them into their product type, it can give exact dimensions where in the real time video or image that product is located. This script is implemented into the robotic system which then picks the product detected and send it to the place where it can processed to recycled or make fuel. The project increases life expectancy of humans and animals of near-by areas. Government spend lot of money behind those manual scavengers for their health and for picking this products, which can reduce. Landfill Refinery prevents from hazardous and brutal conditions that occur due gases and fires caused by this landfill (large dump area). Various training methods and ML algorithm were practiced/ tested to improved the accuracy of results. Please see the video below for this solution.
Sanskrit NLP is a voice Enabled Sanskrit Speech Recogniton using machine learning. When a user says a common Sanskrit phase, the program shall be able to recognize the Sanskrit word using language acoustic model, correct the grammar errors using ML classification algorithm.
Whenever someone wants to create a pdf file or document file of the book which is written in Sanskrit it's quite difficult to convert into a document and there is also a chance of Sanskrit grammar mistake, using this one can easily convert in a document file.
Auto Sentiments of Social Media
This solution allows any company/entity to track the popularity (or not) of their product/solution/campaign based on Social Media messages (i.e. Tweets). This python and NLTK based solution leverage data analytics capabilities of Python (numpy and Matplotlib) and can draw runtime charts of sentiments as they are being expressed at scale.