Cerberus-Dog-Breed-Classification-and-Body-Localization-PyTorch!

The goal of the project is to develop a PyTorch-based model capable of recognizing and classifying 3 dog breeds and localizing 3 specific body parts within images.


View Demo · Report Bug · Request Feature

(back to top)

1. Just Clone/Download the repo

(back to top)

Classification and Localization functionalities are available in their respective user interfaces (GUIs).

To utilize the classification feature, please follow these steps:

• Install the required dependencies mentioned in the .txt file.
• Run the .\main.py file.
• Within the classification GUI, you have the option to create a new alexnet model from scratch. This model can then be used for training, validation, or prediction purposes. If you already possess a pre-existing model, you can select it within the interface. Following the training phase, you can evaluate the model's performance using a cat dataset and examine the results.

If you wish to utilize the localization feature, please run the .\Object-detection\main.py file. (Note: A unified GUI will be introduced in a future version of the project.)

• In the localization GUI, you can upload an image of a dog and the model. This allows you to precisely identify and localize various body parts.
• Additionally, you can monitor the output in real-time during this process by selecting the "Show model validation data" option.

(back to top)

Mandatory features:

• Recognizing and classifying 3 dog breeds (Toy Poodle, Beagle, Siberian Husky)
• Localizing 3 specific body parts (eyes, nose, tail)

Optional features:

• GUI

(back to top)

Distributed under the MIT License. See LICENSE for more information.

(back to top)

Gerardo Cipriano - gerardo.cipriano@studio.unibo.it

Massimiliano Battelli - massimilian.battelli@studio.unibo.it

Project Link: https://github.com/gerardocipriano/Cerberus-Dog-Breed-Classification-and-Body-Localization-PyTorch

(back to top)