Ranging from inspection systems in the manufacturing industry to vision-based drone guidance systems in precision agriculture, computer vision technology boosts the productivity of your processes and the quality of your products.
Our team has built custom computer vision solutions with advanced components such as object classification, feature recognition, segmentation, pattern recognition, object detection, filtering etc. to address the business challenges of diverse industries.
Identify people in an image, from face detection and processing.
Detect objects in an image such as people, cars, chairs, traffic lights, etc.
Identify the scene in an image like restaurants, stadiums, markets etc.
Extract text from simple photos or photos with more complex backgrounds.
Femsa wanted us to create a computer vision algorithm that was able to detect different bottle types. The objective of this project was to count the number of bottles as well as detect if a bottle was misplaced on the factory line.
Small dataset: using transfer learning on top of existing models we were able to reach first-in class results.
Create a search engine for images and videos.
Made more than 15 custom AI models – most of them non-existent previously – to solve problems like face recognition, soccer play recognition, film elements recognition, logo recognition and more.
Optimize cargo transportation. Prevent wood transport trucks from reaching the scales without having the appropriate load.
Security cameras were installed at the entrance of the scales where the trucks weight their cargo. Using computer vision, it began to automatically check if the truck was loaded enough to enter.
USD 800K savings were obtained by implementing this solution in the first few months.
Pyxis helps companies and individuals to collect, organize and explore data to make smarter decisions. Data Science, Big Data, Machine Learning, Computer Vision, Natural Language Processing and Data Engineering.
We are here to help you!
You can leave us your query or recommendation through this form.