SKYLAR

AI Vision Platform

Computer Vision is a sub-field in the larger realm of Artificial Intelligence (AI) that allows the digital systems to process, analyze and derive meaningful information from visual data sources (images or videos) and take necessary actions. The most common branches of computer vision include Object Detection, Object Identification, Object Tracking, among which Object Classification is one of the most extensively utilized tasks. The underlying principle behind object classification is to classify the object in a visual content to one of the defined categories. There are several areas where image classification can be applied such as Healthcare, Education, Agriculture, Manufacturing, Inspection and Monitoring, Defense and Security etc.

In the recent years, with the evolution of deep learning models, the availability of computational resources and the accessibility to huge data sources, the accuracy and efficiency of computer vision tasks have increased exponentially. However, developing these computer vision based intelligent systems requires human resources with substantial technical expertise. There is a huge cost involved in managing the skilled human manpower and computational resources. Overall, there is a lack of a system that allows the business organizations to harness the capabilities of intelligent systems for computer vision tasks with minimal technical involvement.

SKYLAR is a noble platform that allows any business organizations or enterprises from any domain to build, train, manage and use any image classification model with minimal technical jargon and easy-to-use interface. It is a one of a kind Vision-as-a-Service (VaaS) model for computer vision tasks, specifically image classification. The core features of the system are

  • Build any number of image classification models based on the datasets within TWO simple steps with no technical experience.
  • The system automatically selects the optimal hyperparameters to achieve an accuracy of >90% for all the classification models with faster training and inference time.
  • Easily re-train an existing model with additional images to improve the accuracy.
  • Generate a prediction API for each classification model that can be embedded within other internal business systems.
  • Automatically manage the computational resources based on the task in hand.
  • Easy-to-use interface to track and manage the image classification models.

The business advantages of SKYLAR include

  • Cost effective – Reduces the cost of human and computational resources
  • User Friendly Interface
  • Time management – Mitigate the time factor required for development of the model
  • Easy Integration – Integrate the model inference APIs into other applications smoothly

Use Cases

01

Medical Diagnosis in Healthcare

02

Real Esate and Retail

03

Manufacturing - Inspection and Defects Identification

04

Governments-Inspection and Monitoring

05

Apparel - Inspection and Defects Identification

06

House Inspection