Finetuned-ViT-Base Image Orientation Detection Model - Open-source and Free to Identify Whether an Image is Inverted

Finetuned Vit Base Patch16 224 Upside Down Detector

Developed by jaygala24

An image orientation detection model fine-tuned from ViT-base-patch16-224-in21k, used to identify whether an image is upside down

Downloads 37

Release Time : 4/2/2022

Model Overview

This model can detect whether an image is upside down, achieving 89.47% accuracy on a custom dataset.

High Accuracy

Achieves 89.47% accuracy on the test set

Based on ViT Architecture

Utilizes Vision Transformer architecture with powerful image feature extraction capabilities

Fine-tuned Model

Fine-tuned from a pre-trained model to adapt to specific image orientation detection tasks

Image Orientation Detection

Binary Classification Task Processing

Image Feature Extraction

Image Processing

Image Orientation Correction

Automatically detects and corrects upside-down images

89.47% detection accuracy

Image Quality Inspection

Automatically identifies incorrectly oriented images in image processing pipelines

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base