DeepHawkeye is a library to detect unusual patterns in images using features from pretrained neural networks

Last update: Dec 13, 2022

Overview

English | 简体中文

Introduction

DeepHawkeye is a library to detect unusual patterns in images using features from pretrained neural networks

Reference PatchCore anomaly detection model

Major features

Using nominal (non-defective) example images only
Faiss(CPU/GPU)
TensorRT Deployment

Installation

$ git clone https://github.com/tbcvContributor/DeepHawkeye.git
$ pip install opencv-python
$ pip install scipy

# pytorch
$ pip install torch==1.8.0+cu111 torchvision==0.9.0+cu111 torchaudio==0.8.0 -f https://download.pytorch.org/whl/torch_stable.html


#install faiss
# CPU-only version（currently available on Linux, OSX, and Windows）
$ conda install -c pytorch faiss-cpu
# GPU(+CPU) version （containing both CPU and GPU indices, is available on Linux systems）
$ conda install -c pytorch faiss-gpu
# or for a specific CUDA version
$ conda install -c pytorch faiss-gpu cudatoolkit=10.2 # for CUDA 10.2

Checkpoints and Demo data

Wide ResNet-50-2 and demo data

[Google]

[Baidu],code:a14e

${ROOT}
   └——————weights
           └——————wide_r50_2.pth
   └——————demo_data
           └——————grid
                    └——————normal_data
                    └——————test_data
           └——————....

Demo

bulid normal lib
python demo_train.py -d ./demo_data/grid/normal_data -c grid
pytorch infer
python demo_test.py -d ./demo_data/grid/test_data -c grid
tensorrt infer
python demo_trt.py -d ./demo_data/grid/test_data -c grid -t ./weights/w_res_50.trt

Tutorials

Need normal example images to cover all scenarios as much as possible
Faiss Documentation Default IVFXX, PQ16

train args

def get_train_args():
    parser = argparse.ArgumentParser()
    parser.add_argument('-d','--total_img_paths',type=str, default=None)
    parser.add_argument('-c','--category',type=str, default=None)
    parser.add_argument('--batch_size', default=64)
    parser.add_argument('--embedding_layers',choices=['1_2', '2_3'], default='2_3')
    parser.add_argument('--input_size', default=(224, 224))
    parser.add_argument('--weight_path', default='./weights/wide_r50_2.pth')
    parser.add_argument('--normal_feature_save_path', default=f"./index_lib")
    parser.add_argument('--model_device', default="cuda:0")
    parser.add_argument('--max_cluster_image_num', default=1000,help='depend on CPU memory, more than total images number')
    parser.add_argument('--index_build_device', default=-1,help='CPU:-1 ,GPU number eg: 0, 1, 2 (only on Linux)')

tips：

--input_size: trade off between speed and accuracy of the result --max_cluster_image_num：If RAM allows, greater than or equal to the total number of samples

test args

def get_test_args():
    parser = argparse.ArgumentParser()
    parser.add_argument('-d', '--test_path', type=str, default=None)
    parser.add_argument('-c', '--category', type=str, default=None)
    parser.add_argument('--model_device', default="cuda:0")
    parser.add_argument('--test_batch_size', default=64)
    parser.add_argument('--embedding_layers', choices=['1_2', '2_3'], default='2_3')
    parser.add_argument('--input_size', default=(224, 224))
    parser.add_argument('--test_GPU', default=-1, help='CPU:-1,'
                                                       'GPU: num eg: 0, 1, 2'
                                                       'multi_GPUs:[0,1,...]')
    parser.add_argument('--save_heat_map_image', default=True)
    parser.add_argument('--heatmap_save_path',
                        default=fr'./results', help='heatmap save path')
    parser.add_argument('--threshold', default=2)
    parser.add_argument('--nprobe', default=10)
    parser.add_argument('--n_neighbors', type=int, default=5)
    parser.add_argument('--weight_path', default='./weights/wide_r50_2.pth')
    parser.add_argument('--normal_feature_save_path', default=f"./index_lib")

tips：

--threshold: depend on scores of anomaly data

result format：{filename}_{score}.jpg

TensorRT Deployment

License

This project is released under the Apache 2.0 license.

Code Reference

https://github.com/hcw-00/PatchCore_anomaly_detection embedding concat function : https://github.com/xiahaifeng1995/PaDiM-Anomaly-Detection-Localization-master

DeepHawkeye is a library to detect unusual patterns in images using features from pretrained neural networks

Related tags

Overview

Introduction

Installation

Checkpoints and Demo data

Demo

Tutorials

License

Code Reference

Owner

CV Newbie

Blender scripts for computing geodesic distance

Demo for the paper "Overlap-aware low-latency online speaker diarization based on end-to-end local segmentation"

PyTorch Personal Trainer: My framework for deep learning experiments

Source code for the paper "PLOME: Pre-training with Misspelled Knowledge for Chinese Spelling Correction" in ACL2021

Neural network graphs and training metrics for PyTorch, Tensorflow, and Keras.

[CVPR 2021] NormalFusion: Real-Time Acquisition of Surface Normals for High-Resolution RGB-D Scanning

Python package for visualizing the loss landscape of parameterized quantum algorithms.

The code of Zero-shot learning for low-light image enhancement based on dual iteration

ObsPy: A Python Toolbox for seismology/seismological observatories.

One-line your code easily but still with the fun of doing so!

PyTorch implementations of the paper: "DR.VIC: Decomposition and Reasoning for Video Individual Counting, CVPR, 2022"

NHS AI Lab Skunkworks project: Long Stayer Risk Stratification

GANSketchingJittor - Implementation of Sketch Your Own GAN in Jittor

VIsually-Pivoted Audio and(N) Text

Text Generation by Learning from Demonstrations

Human head pose estimation using Keras over TensorFlow.

A neuroanatomy-based augmented reality experience powered by computer vision. Features 3D visuals of the Atlas Brain Map slices.

TEA: A Sequential Recommendation Framework via Temporally Evolving Aggregations

Final project code: Implementing BicycleGAN, for CIS680 FA21 at University of Pennsylvania

A robust camera and Lidar fusion based velocity estimator to undistort the pointcloud.