iNet-API-Demo (Demo program for iNet-API)

• This is a repository for demo of iNet-API developped by AIPro Inc.
  • iNet: AIPro Deep Learning Solution supporting the following functionalities:
    • Object Detection and Tracking (RGB and Thermal)
    • Pedestrian Attribute Recognition
    • Pedestrian Counting & Zone Hitmap Estimation
    • Fire Classification
    • Crowd Counting

Licensing and Restrictions

• Commercial use must be approved by AIPro Inc.
• The maximum number of inferences for all channels is limited.
• (Important!!) there are several hidden features to prevent illegal use in this repository.

---

Dependency

• cuda 12.9.0 (cuda_12.9.0_576.02_windows.exe)
• cuDNN 9.12.0 (cudnn_9.12.0_windows.exe)
• openvino 2023.2.0

Installation and Solution Guide

• You can refer to one of the following technical documents in the repository for installation and guide:
  • Korean: AIPro_iNet_solution_guide_v1.4.8(Korean).pdf

Download and extract files

• Download and upzip one of the followings zip files. Then, copy and paste bin, inputs, and videos directories to the solution directory (the directory including the .sln file):
  • Cuda compute capability of your GPU should be 8.9(RTX-40xx) or 12.0(RTX-50xx):
    • RTX-40xx: https://drive.google.com/file/d/1HJFLXkGO9u3tty5ZBPaV7LbC3XBkiKNO/view?usp=sharing
    • RTX-50xx: https://drive.google.com/file/d/1EJ-3EBxbpE_o-RCCa0mNi3hf4Ydj70K-/view?usp=sharing

Run the project

• Open the following sln:
  • iNet-API-Demo.sln in the solution directory