Can AWS provide Real-Time Edge Object Detection Model?


We are in need of an object detection model suitable for deployment on an edge device, specifically the Oak-D. Our primary focus is on the following object classes:

  • Pedestrian
  • Cyclist
  • Motorcyclist
  • Car
  • Bus
  • Truck

However, if these specific labels are not available, an alternative set of labels could be considered:

  1. Person
  2. Bicycle
  3. Motorcycle
  4. Car
  5. Bus
  6. Truck

It's crucial to note that we are not seeking an API solution due to real-time requirements that may introduce latency. Instead, we are interested in a standalone model file that can be deployed in urban and suburban areas, installed on a pole for object detection and tracking purposes.

Additionally, we would appreciate information on the pricing for the service. Please provide details on the cost associated with acquiring and deploying the model file for the specified object detection requirements.

1 Answer

Hey Amun, Rekognition doesn't support on-prem deployment or inference. We had a similar problem before, and ended up building DirectAI ( to solve this.

We let you create custom labels to do object detection & classification, all with zero training. We're actively working on building on-prem solutions for edge devices right now. We'd love to chat if any of this sounds interesting -- here's our Calendly. Hoping to hear how we can meet your needs.

answered 7 months ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions