resnet34

使用TensorFlow自己搭建一些经典的CNN模型，并使用统一的数据来测试效果。

Few-shot learning experiments mostly on speaker recognition.

This repository contains Final project of CSE428 Brac University

Explainable Speaker Recognition

End-to-end Deep Learning & Computer Vision pipeline to harvest, classify, and sort NASA & MAST astronomical imagery using PyTorch (ResNet-34) with SQLite logging and a Flask dashboard.

A PyTorch Deep Learning pipeline utilizing a U-Net (ResNet34 backbone) to perform semantic segmentation of post-disaster satellite imagery for building damage mapping.

Image Segmentation using Oxford Pets Dataset with the goal to improve animal tranquilizer aiming system.

U-Net segmentation algorithm with options of pretrained resnet34 and resnet50 encoders. All of the project dockerized with gpu suppport on anaconda environment with multiple loss support..

📖 Research Overview : Deep learning-based image tampering detection & localization using UNet + pretrained ResNet-34 with Multi-Quality ELA. Built via a 60+ experiment ablation study on CASIA v2.0.

AI-powered web application for real-time plant leaf disease detection using a fine-tuned ResNet-34 CNN, built with PyTorch and Flask.

Diabetic retinopathy severity grading on the APTOS 2019 fundus image dataset - ResNet34 baseline and ConvNeXt V2 (FCMAE-pretrained) PyTorch pipelines.

Multimodal Deep Learning System for Amazon Product Price Prediction using Sentence Transformers and ResNet34

BD-celebrity-face-recognition

Top 5% on Kaggle leaderboard using fast.ai library and resnet50 along with transfer learning.

Detecting Action performed in a video using resnet34 for spatial and temporal stream

This is an implementation of ResNet using keras.

End-to-end AI-powered Infant Cry Diagnostic System using PyTorch, ResNet34, YAMNet, FastAPI, Docker, and AWS S3 with a live inference API.

Diabetic retinopathy severity classifier — ResNet34 fine-tuned for 5-class fundus grading (0.815 self-reported val accuracy). fastai training notebook + ONNX weights.

A semantic segmentation project using U-Net architecture with a simple from scratch one and a pretrained ResNet34 encoder to segment cats and dogs from background.