Automotive Company
Framework For Automated Computer Vision Model Evaluation
The client was a leading automotive technology company, specializing in pioneering products and services for sustainable and connected mobility of people and goods. Established in the 19th century, they now provide safe, efficient, intelligent and cost-effective offerings for vehicles, machines, traffic and transportation. Their endeavors to develop cutting-edge innovations, such as self-driving cars, necessitate evaluation frameworks—one of which Hiflylabs helped create.
Automation
saves engineers days of work per every iteration.
Challenge
Our client leverages multiple deep learning computer vision models for object detection and image classification purposes. Performance evaluation of such self-driving car models is a vital issue, due to their mission critical, multi-purpose and complex nature.
Solution
Hiflylabs provided an automated evaluation framework and dashboarding system. This system handles multiple deep learning computer vision models, different scenarios, weather & road conditions, and objects—such as cars, motorcycles, pedestrians—together with multiple objectives and statistical metrics. By using the new evaluation framework, the client’s engineers can save days on each iteration of their own models, and have a standardized, automated, more detailed and holistic view of their processes than they had before. This enabled deep learning engineers and architects to focus more of their time on development, research and efficient modeling.
Services
Complex Correlation Analysis
Classification
Bootstrapping
Deep Learning Evaluation
System Automation
Industries
Automotive
Technologies
Python
-
