Skip to content

MirazulHasan/FOOD_IMAGE_SEGMENTATION

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 

Repository files navigation

Food Image Binary Segmentation with UNET and MobileNet

This project implements a binary segmentation model to detect the presence of food in images. The model combines the strengths of UNET and MobileNet for efficient and accurate segmentation.

Dataset

Images: 50 food-related images curated for this project.

Annotations: Binary masks were manually created using Photoshop to label areas containing food.

Model Architecture

Backbone: MobileNet serves as the lightweight feature extractor, providing efficiency and flexibility.

Segmentation Head: UNET architecture is employed to output binary segmentation masks.

Workflow

Data Preparation:

Images were manually masked to distinguish food regions (foreground) from non-food regions (background).

Preprocessing steps include resizing, normalization, and augmentation.

Model Training:

The UNET-MobileNet architecture was trained to produce binary masks indicating the presence of food.

Evaluation:

Performance metrics such as accuracy, IoU (Intersection over Union), and Dice Coefficient were used to validate the model.

Features

Binary Segmentation: Detects whether food is present in an image with pixel-level accuracy.

Manual Annotations: High-quality ground truth masks created with Photoshop.

Lightweight Design: Utilizes MobileNet for efficient inference on resource-constrained devices.

Applications

Useful for automated food identification systems.

Can serve as a foundation for further classification or calorie estimation systems.

Explore the repository to see the code, training details, and segmented results in action!

About

This project implements a binary segmentation model to detect the presence of food in images. The model combines the strengths of UNET and MobileNet for efficient and accurate segmentation.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors