Object-relationship graph

Building object-relationship graph from image

Hours total

1640

Technical Stack

PCA, K-means, DTW, Bayesian approach, Cart, VAE, SVM

Services Involved

Dedicated Team

Have similar project?

OVERVIEW

The core task in many image processing and computer vision applications is scene recognition, from search by image to self- driving cars. State-of-the-art machine learning algorithms solve such tasks, although, TensorFlow framework pensures easier and faster implementation of these algorithms.

CHALLENGE

To build a graph of objects and relationships from the pixels of a certain image. The task implies locating objects, recognizing them, finding and classifying relationships between the objects.

Solution

End-to-end learning is a state-of-the-art approach for image analysis and other machine learning tasks. Instead of stacking several models learned independently of each other, end-to-end learning provides a faster way to do all the optimizations by a single training run. The model was tested on the VisualGenome dataset and had several outputs: objects' positions, their classes, relationships between objects and types of relationships.

Results

End-to-end model capable to build an object-relationship graph from the image pixel data without using any intermediate models.

Recent cases

Our latest challenges where technology meets creativity

image
image

Clusterization

The technique of finding similarities in the data point and grouping similar entities together

image
image

Compare Telco

Know more. Buy smarter

image
image

Costed Engine

Get images, adjust damage repair prices, generate reports, send data

Contact us

If you have a question, request or just want to meet up for coffee, call, email us or fill out the form and we will contact you asap.

Full Name
Email
Short message