In this tutorial, you will perform inference across 10 well-known pre-trained object detectors and fine-tune on a custom dataset. Design and train your own object detector.

A step-by-step PyTorch guide for improved vehicle type 2D object detection based on Faster-RCNN using transfer learning and data augmentation.

Here I have used MobileNet-SSD, deep learning objection technique, for detecting various objects in the UC-Win/Road virutal reality based enviroment. The results of the standard MobileNet-SSD detector was not good enough. Therefore, Transfer Learning Technique is used for fine tuning of the model. Then the new model is applied for detection of the UC-Win/Road objects.

This tutorial guides you through the process of creating a custom object detector for cat eyes using YOLOv11 Nano from Ultralytics. We’ll cover the entire pipeline — from collecting and labeling data to fine-tuning a pre-trained YOLOv11 Nano model for cat eye detection.

Step-by-step guide for deployment of an object detector model on an MCU-based system. The object detector model is developed by fine-tuning a MobileNet model using transfer learning.

Zero-shot object detection is supported by the OWL-ViT model which uses a different approach. OWL-ViT is an open-vocabulary object detector. It means that it can detect objects in images based on free-text queries without the need to fine-tune the model on labeled datasets.

In this project I've made a Drive Drowsiness Detector using Python, Pytorch and Yolo Model for Object Detection. Our aim was to build a fine tuned model which can help us to determine whether the driver is feeling drowsy or awake. We have used Ultralytics Yolo(v5) in this and used its base model which is trained on Coco dataset for object detection. After all the computation we were able to detect the face and provide the desired output in the real time by leveraging our webcam.

Evaluating and iterating on MBARI's fine-tuned benthic object detector

An object detector trained with a Kaggle GPU, using Tensorflow and a fine-tuned ssd resnet50.

A production-ready drone detection system demonstrating transfer learning, fine-tuning, and modern MLOps practices for AI/ML deployment. The system leverages transfer learning to adapt a general-purpose object detector to the specialized task of drone detection.

The backend of RecycloVision, an AI-powered recycling assistant that detects waste materials in images and returns real-world carbon impact insights. Built with Flask and PyTorch, it uses a YOLOv8 object detector and a fine-tuned ResNet classifier to analyze user-uploaded images and provide accurate recycling information.

Repository for fine-tuning object detectors

Fine tuning 2024 SOTA object detector on a traffic light dataset "Lisa"

Object detection for HIOD dataset (inside Hospitals) using fine-tuned benchmarked detectors.

Vision Transformer (Conditional DETR) based object detector fine-tuned on custom drone dataset.

Object detection using YOLO-V3 object detector alongside deep learning concepts such as image classification, testing and fine tuning implemented.

Fine tuning object detectors for deep sea environments and adapting them to be deployed in underwater vehicles.

Fine-tuned YOLOv8 object detector for agricultural imagery (Global Wheat Detection) with full data pipeline, experiment tracking and performance evaluation.

🍬 Candy Counter is a trained object detector which counts different types of candy. Created by annotating images of candy In Label Studio and fine-tuning an object detector by adapting a Huggingface tutorial.

Sign Language Detector using TensorFlow Object Detection API. It collects images of hand gestures and fine-tunes a pre-trained object detection model to detect common sign language gestures in real-time.

Comparative study of YOLOv8, YOLOv11, YOLOv12 for object detection, fine-tuning YOLOv8 for face detection on WIDER FACE, benchmarking with YOLOv8-Face, and training a YOLOv1-style face detector.

Transfer-learning YOLO pipeline: auto-annotate Flickr30K with Florence-2, then fine-tune a lightweight YOLO-nano detector on a balanced pet/person dataset for high-precision object detection

Implementation of a U-Net with VGG16 backbone for brain tumor segmentation from MRI scans, and fine-tuning of both two-stage (Faster R-CNN) and single-stage (SSD) object detectors for traffic sign recognition.

This project uses PyTorch's vision references to create a pipeline for training an object detector for chess pieces. The model is fine-tuned from a pre-trained Faster R-CNN, initially trained on the COCO dataset.

Project for automated analysis of PDF documents to detect signatures, seals, and qr codes. The system converts PDFs to images, runs a YOLO-based object detector fine-tuned on document data, and outputs bounding boxes for every detected item

Bournemouth University Campus Pathfinder: A CV Solution for Building Detection & Classification This project develops and benchmarks deep learning models to detect and classify key buildings on Bournemouth University's Talbot Campus. It features a comparative analysis of a custom ResNet-based object detector and a fine-tuned YOLOv5 model