https://github.com/mehrab-kalantari/cat-landmarks-estimation

Predicting cats feature points using a feature pyramid network
https://github.com/mehrab-kalantari/cat-landmarks-estimation

feature-pyramid-network fpn mobilenetv2 neural-network-regression

Last synced: 5 days ago
JSON representation

Predicting cats feature points using a feature pyramid network

• Host: GitHub
• URL: https://github.com/mehrab-kalantari/cat-landmarks-estimation
• Owner: Mehrab-Kalantari
• Created: 2024-07-05T14:00:57.000Z (5 months ago)
• Default Branch: main
• Last Pushed: 2024-07-05T14:09:45.000Z (5 months ago)
• Last Synced: 2024-10-24T15:52:58.204Z (27 days ago)
• Topics: feature-pyramid-network, fpn, mobilenetv2, neural-network-regression
• Language: Jupyter Notebook
• Homepage:
• Size: 1.37 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Cats Feature Points Estimation

Predicting cats feature points using a feature pyramid network

[Kaggle Dataset](https://www.kaggle.com/datasets/crawford/cat-dataset)

## Understanding Data

Each record contains an image and a file with 19 values.

The first value represents the number of points (which is 9)

Each two next values represent a coordinate in the picture which shows a feature point:

* Left Eye

* Right Eye

* Mouth

* Left Ear 1

* Left Ear 2

* Left Ear 3

* Right Ear 1

* Right Ear 2

* Right Ear 3

* ![pic](sample/sample.png)

**Each image has been resized to 224 * 224**

## Modeling

* MobileNetV2 as feature extractor

* Feature pyramid with 4 levels

* Detection head with 18 outputs (each corresponding to a point)

## Evaluation

* ![eval](sample/loss.png)