Aim of the Project:
The project aims to develop an advanced system using Deep Learning to identify dog breeds from images accurately. The goal extends beyond merely recognizing and classifying over various distinct breeds; it also includes discerning service dogs and domestic dogs adorned with clothing. This classification acknowledges the unique roles of service dogs, necessitating their quick identification in diverse environments, thereby enhancing accessibility and support for individuals with disabilities. Additionally, recognizing domestic dogs in clothing caters to the practical aspects of pet ownership where dogs are often dressed for weather protection or fashion.
This comprehensive approach aims to broaden the utility of the system across multiple platforms, including pet adoption agencies, veterinary clinics, and public safety measures.
Neural Network/ML Techniques:
- ResNet50
- Fully connected neural network
- YOLOv5
- Data Augmentation
- Feature Extraction
- Deployed Application: Dog Breed Classifier https://huggingface.co/spaces/marciayang/dog-breed-detect
- Application on video