Introducing “VehicleAttributeRecognitionBarrier,” a machine learning model available for use with ailia SDK. By utilizing the ailia SDK and ailia MODELS, an edge-oriented inference framework, it becomes easy to implement AI functionality into applications.
Overview of VehicleAttributeRecognitionBarrier
VehicleAttributeRecognitionBarrier is a machine learning model developed by Intel for identifying vehicle attributes. It can detect the type and color of vehicles.
open_model_zoo/README.md at master · openvinotoolkit/open_model_zoo
This model presents a vehicle attributes classification algorithm for a traffic analysis scenario. Color average…
Architecture of VehicleAttributeRecognitionBarrier:
VehicleAttributeRecognitionBarrier takes input as frontal images of vehicles and outputs their attributes. The detectable attributes include color and type of the vehicle.
The color category consists of seven options, while the vehicle type category consists of four options. The accuracy rates are 82.71% for color and 87.34% for vehicle type.
The model architecture features a ResNet-like structure. The input is an image of size (1,72,72,3), and the output consists of a vector of size (1,7) indicating the probabilities of colors and a vector of size (1,4) indicating the types of vehicles.
As a constraint, the model requires front-facing images of vehicles, and the occlusion should be less than 50%.
Usage of VehicleAttributeRecognitionBarrier
To use VehicleAttributeRecognitionBarrier, use the following command: first, detect cars in any video using YOLOv3, then proceed to detect their attributes.
$ python3 vehicle-attributes-recognition-barrier.py -v input.mp4
Here is an example of usage.
ailia-models/vehicle_recognition/vehicle-attributes-recognition-barrier at master ·…
(Image from…
Ax Inc. is developing ailia SDK, a cross-platform AI inference framework that utilizes GPUs for high-speed processing. As a company focused on practical AI applications, Ax Inc. offers a comprehensive solution for AI, including consulting, model development, SDK provision, AI-based app/system development, and support. Feel free to reach out to us for any inquiries related to AI.
