marco-willi / camera-trap-classifier

Automatically identify animals in camera trap images by training a deep neural network.
MIT License
49 stars 17 forks source link

Code for Windows #22

Closed sazidabintaislam closed 5 years ago

sazidabintaislam commented 5 years ago

I am working on Camera Trap Images. I have just started and trying to understand the codes. It would be great if I could get an idea or resources for windows.

marco-willi commented 5 years ago

From what I have read, it is difficult to train models on Windows. The reason for this is the need to employ a GPU to effectively train models. Therefore, if your goal is to train a model on your own dataset I would recommend using cloud computing ressources. If you are not familiar with how to use cloud computing ressources I suggest to follow this guidelines Link and/or ask someone with Linux/Unix experience.

However, if your goal is to apply trained models, Windows is perfectly fine. I recommend to:

  1. Install the Anaconda distribution Link
  2. Follow the guidelines to use trained models here: Link
  3. Follow the instructions here to apply the model: Link
sazidabintaislam commented 5 years ago

I am doing a project where I will detect Frog from images. I have collected 200 images but those are not labeled. Can you suggest me a good data labeling tool.

marco-willi commented 5 years ago

I can't recommend a specific tool. There are many available and it depends a lot on what you feel comfortable with and what the output of these tools is.

200 images does not sound like a lot if it is the total number of images (and not per species). If the goal is to detect the presence of a frog you may want to check out this pre-trained object detection model: https://github.com/microsoft/CameraTraps