Use Case Title
Cat and Dog Detection – Real-Time Pet Detection using YOLOv9
Team Members
Drashti Shetty
Shruti Thakur
Kedar Phadke
Nikhil Dubey
Krunal Garud
Use Case Importance
Accurate animal detection systems can be useful in smart surveillance, pet monitoring, veterinary assistance, and automated image classification systems. Our solution helps in detecting cats and dogs in real time with improved accuracy and efficient object recognition.
Data Collection and Annotation
We collected and prepared a dataset containing images of cats and dogs from various real-world environments under different lighting and background conditions.
Classes annotated:
- Cat
- Dog
Images were labeled and split into training, validation, and test datasets.After augmentation total images in:
Train:5205 Images
Valid:495 Images
Test:248 Images
Sample Image from dataset:
Model Training and Validation
We trained the YOLOv9 model using multiple training experiments to improve real-time detection performance and classification accuracy.
Metrics monitored:
- Precision
- Recall
- Mean Average Precision (mAP)
- Training and Validation Loss
The confusion matrix showed high correct predictions with low misclassification rates. Precision, Recall, and mAP graphs demonstrated stable learning and improved detection performance over training epochs.
Model Deployment and Demo video
The trained model was tested for real-time inference and object detection performance.
Performance observations:
- Accurate cat and dog detection in most conditions
- Stable bounding box predictions
- Good performance under varying lighting conditions
- Efficient real-time detection capability
The model demonstrated reliable detection results with smooth inference performance.
Demo Video Link: https://drive.google.com/file/d/1fz55jEOmfuS1ojMPrZzJDLSNov1xMNKe/view?usp=drivesdk
Conclusion
Our YOLOv9-based Cat and Dog Detection model achieved strong detection accuracy with stable learning performance during training and validation.
Through this project, we learned the importance of:
- Proper dataset preparation
- Accurate annotation
- Data augmentation techniques
- Model evaluation using Precision, Recall, and mAP
- Real-time AI deployment and optimization.
This project provided valuable hands-on experience in Computer Vision, Object Detection, and Deep Learning applications.
Special Thanks to :
@bothe sir, Sandeep Dwivedi sir for providing this opportunity for exploring and learning new concepts.