tags: - image-segmentation

Custom U-Net Model for Semantic Segmentation

Overview

This repository contains a custom U-Net model for semantic segmentation tasks using TensorFlow/Keras. The model is designed to perform pixel-wise classification on images, such as distinguishing between different classes in medical images or segmenting objects in natural images.

Model Architecture

The model architecture is a standard U-Net with custom convolutional blocks in the encoder and decoder. The encoder consists of four downsampling blocks, and the decoder consists of four upsampling blocks. The bottleneck layer in the middle has a larger number of filters.

Encoder

  • Blocks: 4
  • Filters: [64, 128, 256, 512]

Bottleneck

  • Filters: 1024

Decoder

  • Blocks: 4
  • Filters: [512, 256, 128, 64]

Training Details

  • Loss Function: Categorical Crossentropy
  • Optimizer: Adam
  • Learning Rate: 0.001
  • Batch Size: 16
  • Epochs: 50
  • Validation Split: 20%

Usage

Installation

Ensure you have the necessary dependencies installed. You can install them using pip:

pip install tensorflow
Downloads last month
18
Inference Providers NEW
This model is not currently available via any of the supported third-party Inference Providers, and HF Inference API was unable to determine this model's library.