How to Use SAHI Tiling Windows to Detect Small Objects with YOLOv8 or ANY Object Detection Model

Поделиться
HTML-код
  • Опубликовано: 10 янв 2025

Комментарии • 11

  • @NicolaiAI
    @NicolaiAI  8 месяцев назад +1

    Join My AI Career Program
    👉 www.nicolai-nielsen.com/aicareer
    Enroll in My School and Technical Courses
    👉 www.nicos-school.com

  • @jaidityakapoor6801
    @jaidityakapoor6801 4 месяца назад

    Can we do this on a custom dataset ?

  • @alihan_ozturk
    @alihan_ozturk 7 месяцев назад

    Why didn't you set the windows as a batch and run them through the model all at once?

  • @RonalRomeroVergel
    @RonalRomeroVergel 8 месяцев назад +1

    is there any way to do that in real time??

    • @NicolaiAI
      @NicolaiAI  8 месяцев назад

      Yes with optimized models and good hardware. I’m still running this at 25 fps with 4-6 tiles on a 4090. You can get good speed up with optimization frameworks

  • @АльбертИванов-ц4х
    @АльбертИванов-ц4х 8 месяцев назад +1

    is it the same as Sahi or different approach ?
    where to find code ?

    • @NicolaiAI
      @NicolaiAI  8 месяцев назад +1

      The underlying code behind Sahi. Same approach. Code is on my GitHub, I have forgotten to include the link..

  • @cyvision
    @cyvision 8 месяцев назад +2

    Awesome can we use this to segment small objects? @NicolaiAI

    • @NicolaiAI
      @NicolaiAI  8 месяцев назад

      Yeah you can use it with any detection model, object detection, segmentation etc

    • @cyvision
      @cyvision 8 месяцев назад

      ​@@NicolaiAI thanks. how about saving annotation .txt file? do u have plan to add this to your code?

    • @illink7
      @illink7 8 месяцев назад

      @@NicolaiAI Hi thank you for this video, I have already researshed dozen of sites but I can't find solution how to combine segmentation model + sahi. Could you provide more information or write me please🙏🙌