top of page

Video Analytics

Computer Vision solution to optimise the operational efficiency of a restaurant chain

problem (1).png

 

The client needs a solution to analyse table allocation, customer inflow and outflow, the time gap taken for serving each customer, tracking abandoned objects and detecting hazardous situations.

lamp.png

 

We proposed a two-phase approach CV based data collection and analytics

Data collection: Development and implementation of a computer vision-based system to collect the following information:

  • Vacant period of a table

  • Lean period of a Bartender

  • The average engagement time period of a table

Video Analytics:

  • Analyzed a bunch of above mentioned similar operational parameters

  • Powered by deep-learning algorithms to track abandoned objects and detect hazardous situations like fire, loud sound etc.

  • Developed and implemented an algorithm to track the customer inflow and outflow

  • Web & mobile based dashboard to handle the statistics, alerts, captured events from the real-time video feeds

 

Frameworks/Technology: Python 3, Tensorflow, YOLO, MySQL, C#, Angular 5, HTML 5, D3 charts

value (1).png

 

Improved operational efficiency and prediction of customer engagement rate

el-Kane

el-Kane

One-stop solution for visually impaired

problem (1).png

 

This is our own product under the hood. When we studied the market, we got to know that no self-learning wearable product was available for the visually impaired. Especially with a dept calculation mechanism. 

lamp.png

A computer vision-based virtual assistant, which consists of spectacles and earphones

  • Analyse and understand the context of a real-world scenario

  • A user can communicate with the device. Eg- If he/she is in a park and asking for a free bench.

  • Navigation assistance with voice over

  • Alerts during panic scenarios

  • Identify Facebook friends, when they are in the user's vicinity by using facial recognition

  • Social media platform support, to handle the panic or hazardous scenarios

  • Also, trigger alerts to the immediate point of contacts

 

Frameworks/Technology: Python 3, Sphinx, Tensorflow, Machine Learning, FastText

value (1).png

 

State of Art product that can be affordable for every consumer strata

CV solution- Retail

Footfall analysis using headcount and attendance management using facial recognition

Computer Vision Solutions for Retail

problem (1).png

 

A retail chain in the middle east needs a solution to analyse the customer inflow and outflow. They want to automate the attendance management process

lamp.png

  • Bundled up a couple of CV solutions for the retail chain to handle the footfall analysis and attendance management

  • Analyze the inflow and outflow of customers by considering head counts for footfall analysis

  • For the attendance management, less than a minute initial training process is required to recognize the face in the future.

  • A backend dashboard will show the results in numbers and stats.

 

Frameworks/Technology: Python 3, PyTorch, TensorFlow, Power BI

value (1).png

 

Reduced the overhead of attendance management. By implementing CV based footfall analysis, it paved the way for demographic analysis of the customers. The future scope is to track multiple visits, customer emotion etc.

RPA- Back Office

RPA solution for back-office automation

Document template identification and content extraction

problem (1).png

 

The client needs to automate the stereotypic process of filling e-forms, data gathering and entry etc.

lamp.png

  • Developed a CV+NLP based solution that can scan the document, extract contents, understand the context and process accordingly.

  • It also covers the handwriting recognition part

  • Other features are auto form filling, template recognition and populating data accordingly.

 

Frameworks/Technology: Python 3, TensorFlow, Spacy, Gensim, NLTK

value (1).png

 

Save time and reduced the pain of outsourcing and managing an offshore team

Ecommerce Chatbot

Intelligent chatbot for image merchandise e-commerce platform

Computer Vision-based Virtual Assistant

problem (1).png

 

Our client from Germany, a start up firm focused on Image Merchandise was in need of an intelligent solution that can list out images by having a casual chat with users. Also, it should handle search by image feature.

lamp.png

  • We developed a CV based solution by implementing an auto-tagging mechanism

  • Also, it can identify similar images and put out a meaningful recommendation to users

  • No separate web interfaces to upload and search images.

  • While uploading, chatbot will handle auto-tagging by analyzing the image

  • Handled the Payment integration, user profile management (to understand the seller or buyer).

 

Frameworks/Technology: Python 3, Rasa Stack, TensorFlow, NLTK, Angular 5, HTML 5, D3 charts

value (1).png

 

Great user experience, seamless customer interaction and improved sales with the recommender system

RPA- Inside Sales

 

RPA based solution for Inside Sales

Intelligent chatbot for image merchandise platform

problem (1).png

 

Our client was in need of an automate the process of prospecting and lead generation since it consumes a lot of time.

lamp.png

  • Based on the keywords entered by the user it will automate the prospect search process and finally, an excel sheet with prospects will get downloaded

  • The platform will automate the connecting with the prospects through LinkedIn and email

  • If there is no email identified, then it will try various combinations of first and last name and validate the actual email.

  • The user can use preferred email or LinkedIn content to connect with prospects

 

Frameworks/Technology: Python 3, Spacy, NLTK, Selenium, Machine Learning

value (1).png

 

Improved prospect to lead conversion time. Hence the sales activities are in rapid pace now.

Emotion recognition

 

Consumer Feedback Analysis

NLP based solution to analyse the emotion from social media feedback

problem (1).png

 

The client needs a feedback analysis prior to their product launch.

lamp.png

  • We have created a machine-learned model to analyze the user reviews of competitor products by considering 6 human emotions: Anger, Disgust, Fear, Joy, Sadness, and Surprise.

  • We mapped these 6 emotions with the product features and finally delivered a visualization of the same.

 

Frameworks/Technology: Python 3, NLTK, Tensorflow, Gensim, FastText

value (1).png

 

The client knows exactly where they have to stand out when compared to their competitor.

Advanced Analytics
bottom of page