Accurate counting of vehicle axles is essential for traffic control, toll collection, and infrastructure development. We present an end-to-end, video-based pipeline for axle counting that tackles limitations of previous works in dense environments. Our system leverages a combination of YOLO-OBB to detect and categorize vehicles, and YOLO to detect tires. Detected tires are intelligently associated to their respective parent vehicles, enabling accurate axle prediction even in complex scenarios. However, there are a few challenges in detection when it comes to scenarios with longer and occluded vehicles. We mitigate vehicular occlusions and partial detections for longer vehicles by proposing a novel TRAX (Tire and Axle Tracking) Algorithm to successfully track axle-related features between frames. Our method stands out by significantly reducing false positives and improving the accuracy of axle-counting for long vehicles, demonstrating strong robustness in real-world traffic videos. This work represents a significant step toward scalable, AI-driven axle counting systems, paving the way for machine vision to replace legacy roadside infrastructure.

Trax, an end-to-end library for deep learning developed by Google. It is designed to be easy to use, with clear with good speed, with the ability to run on modern hardware such as GPUs and TPUs.The Google Brain team actively uses and maintains Trax. It is built on top of the JAX and TensorFlow numpy, which provides automatic differentiation, a set of numerical operations, and support for GPU acceleration. It includes a wide range of pre-built models and algorithms. In addition to its extensive selection of models and algorithms, it also has bindings to a large number of deep learning datasets, including Tensor2Tensor and TensorFlow datasets. The following code creates a Transformer model for machine translation, initialises it with pre-trained weights, tokenizes an input sentence, decodes the model's output, and then detokenizes the output to get the translation.

"Digitizing" store shelves is a new imperative for Coca-Cola and other CPG companies. In the face of changing consumer behavior and the new retail playbook scripted by Amazon, CPG brands have raced to hop on the "digitization" bandwagon to erase the advantage e-commerce has over brick-and-mortar retail. Now add this area to their get-even list: store shelves. For decades, CPG companies like Coca-Cola have conducted manual audits and surveys to see if, how and where stores stock their goods on shelves, as well as find out if their goods are next to rivals' products and if any out-of-stock items have been replenished. Thanks to technology, this old-school and time-consuming process, prone to human errors, is increasingly being dropped as CPG brands seek to bridge the gap between what they can see in stores and online.

Retailers are looking towards artificial intelligence (AI) such as image recognition technologies to help enhance sales staff productivity as well as make them be more competitive against their peers. Alexander Laugomer, project manager digital merchandising at consumer and industrial goods firm Henkel, says: "The major benefit of the AI technology is that it not only provides us with information on our own and our competitors' products, but it also gives us an actionable report straight to our mobile devices. "This allows our sales reps to improve our brands' situation in-store there and then, without the time-consuming task of manually compiling data." Henkel outsources this function to retail image-recognition firm Trax. The technology works by being able to recognise more than eight million images on a shelf.