A complete data labeling platform which helps machine learning engineers build high quality ground truth datasets for training and validating machine learning models. It breaks down large problems into micro-tasks and distributes among its large community of trained annotators. It works on the principle of microwork, where a series of small tasks which together comprise a large unified project, and are completed by many people over the Internet
Problem
The client provides the annotation platform as well as the human intelligence experts to generate training data for computer vision models at scale. They partner with annotation service providers like Intellekt whenever there is huge quantum of data to be labelled. One of the cases was when the client had to label 80,000 videos on a monthly basis that involved various types of data annotations.
The Intellekt Solution
Intellekt, with its skilled annotation capabilities was keen to take up the challenge. It put together a team of 65 associates and provided Computer Vision data labelling services at scale and with world-class quality. We delivered object annotations by creating bounding boxes around objects, semantic segmentation of images and attributes updations for existing face annotations using their in-house platform.
Result From The Customer
The team was able to fulfil the client’s needs with an accuracy of 99%. The client was indeed delighted to partner with Intellekt as we were able to meet the huge data volumes without compromising on quality or accuracy. The annotations delivered are fed as training datasets for their Machine Learning model to increase its efficiency.
Services Provided
Object Annotation
Semantic Segmentation
Updating Attributes
Video Annotation / Month
%