Upload a photo of any dog here and get a prediction for its breed.
Welcome to my advanced dog breed identifier.
This web application is the culmination of my comprehensive dissertation project, which involved the development of a sophisticated Convolutional Neural Network (CNN) designed to accurately identify dog breeds from single 224x224 photos. The CNN achieves an impressive 85.2% accuracy and has been trained on an extensive dataset comprising 152,000 dog photos, representing 269 distinct breeds.
The training dataset is a meticulously curated collection that encompasses computer vision datasets like Stanford Dogs, Tsinghua Dogs, and Oxford-IIIT Pet., alongside carefully acquired web scraped images. This meticulous approach ensures the model's capability to recognize a diverse range of breeds, regardless of the image source.
One of my primary objectives is to provide an efficient and user-friendly experience. The web app has been optimised for speed and simplicity, allowing users to effortlessly upload a photo of a canine companion and promptly receive the precise breed identification.
For access to the dataset, please contact [email protected].