Skip to content

This is repository containing a full pipeline (from annotation to training) for building an orangutans detector.

Notifications You must be signed in to change notification settings

clementapa/orangutan-image-video-detection

Repository files navigation

Author: Apavou Clément

Python PyTorch YOLOv8

WildGuardian: AI for Orangutan Ecosystem Surveillance 🦧🔍

Image generated using Stable Diffusion XL (dynavisionXL). Prompt: orangutan sitting and working on a computer in the jungle, Disney Pixar movie style, 3D render, good lightning. Negative prompt: (deformed eyes, nose, ears, nose), bad anatomy, ugly, ugly eyes.

👁️ About the model

This is repository containing a full pipeline (from annotation to training) for building an orangutans detector.
The model was trained using this dataset for orangutan images and this dataset as background images. Annotations were obtained using zero shot object detection method GroundingDino.

My experiments are publicly available on Weights & Biases:

You can try the model on HuggingFace Space

Detection results on some videos 🦧.

😁 Good to know

The pipeline can be used to create any type of detector. The pipeline is the following:

  1. Collect images of the target class.
  2. Collect negatives images that does not contain the target class, so that the model generalizes well.
  3. Auto-annotate images of the target class using zero-shot object detector such as GroudingDino (prepare_data/labelize_dataset.py or Open In Colab).
  4. Create empty annotations for negatives images (prepare_data/create_annotation_background_image.py).
  5. Split the datasets (prepare_data/split_data.py).
  6. Create de yaml file containing information on the data to be used during training (data.yaml).
  7. Launch the yolo training (train_yolov8.py) and experiment by tuning training hyperparameters.

If the project you're working on can contribute to wildlife conservation, don't hesitate to contact me, I'll be delighted to help. Also, during my free time, I am willing to work on projects related to wildlife conservation.

🦧 About the orangutans

Because to habitat destruction, illicit poaching, and the pet trade, orangutans are in danger of going extinct. Their natural habitat has been significantly reduced by deforestation and the growth of palm oil plantations. Adult orangutans are occasionally sought for their body parts, and they are frequently captured and sold as pets. Climate change and disease are also taking a toll on their populations. Furthermore, it is concerning to note that they are limited to Borneo and Sumatra, two places on Earth. Sustainable practises and conservation initiatives are crucial to preventing the permanent extinction of these amazing animals.

🌍 AI for good

Artificial Intelligence (AI) has unquestionable power in the realm of innovation and technology. Even though artificial intelligence (AI) has frequently been used for commercial advantage, it is important to stress that AI can also be used for more noble purposes, such as protecting the environment and the planet's future. We can build a more promising and sustainable future if we reorient AI's focus from business to improving our planet.

🗃️ Datasets

🐒 Models released

model training name data train data validation checkpoint Experiment
YOLOv8n train7 1085 images of orangutans & 1095 images of other monkeys 272 images of orangutans & 274 images of other monkeys

📈 Metrics train 7

Validation metrics for the train 7, other metrics are available on the wandb experiment.

About

This is repository containing a full pipeline (from annotation to training) for building an orangutans detector.

Topics

Resources

Stars

Watchers

Forks

Languages