marooncn / Defect Detection Classifier
Visual Defect Detection on Boiler Water Wall Tube Using Small Dataset
Stars: ✭ 91
Programming Languages
python
139335 projects - #7 most used programming language
Labels
Projects that are alternatives of or similar to Defect Detection Classifier
Handwriting Number Classification
A computer vision project, based on cimg library and svm training, to classify handwriting number.
Stars: ✭ 22 (-75.82%)
Mutual labels: svm
Glcm Svm
提取图像的灰度共生矩阵(GLCM),根据GLCM求解图像的概率特征,利用特征训练SVM分类器,对目标分类
Stars: ✭ 48 (-47.25%)
Mutual labels: svm
Gru Svm
[ICMLC 2018] A Neural Network Architecture Combining Gated Recurrent Unit (GRU) and Support Vector Machine (SVM) for Intrusion Detection
Stars: ✭ 76 (-16.48%)
Mutual labels: svm
Ailearning
AiLearning: 机器学习 - MachineLearning - ML、深度学习 - DeepLearning - DL、自然语言处理 NLP
Stars: ✭ 32,316 (+35412.09%)
Mutual labels: svm
Patternrecognition matlab
Feature reduction projections and classifier models are learned by training dataset and applied to classify testing dataset. A few approaches of feature reduction have been compared in this paper: principle component analysis (PCA), linear discriminant analysis (LDA) and their kernel methods (KPCA,KLDA). Correspondingly, a few approaches of classification algorithm are implemented: Support Vector Machine (SVM), Gaussian Quadratic Maximum Likelihood and K-nearest neighbors (KNN) and Gaussian Mixture Model(GMM).
Stars: ✭ 33 (-63.74%)
Mutual labels: svm
Vehicle Detection
Vehicle detection using machine learning and computer vision techniques for Udacity's Self-Driving Car Engineer Nanodegree.
Stars: ✭ 1,093 (+1101.1%)
Mutual labels: svm
Tensorflow cookbook
Code for Tensorflow Machine Learning Cookbook
Stars: ✭ 5,984 (+6475.82%)
Mutual labels: svm
Smart Surveillance System Using Raspberry Pi
This is my Third Year Project for face recognition using OpenCV
Stars: ✭ 41 (-54.95%)
Mutual labels: svm
Sarcasm Detection
Detecting Sarcasm on Twitter using both traditonal machine learning and deep learning techniques.
Stars: ✭ 73 (-19.78%)
Mutual labels: svm
Svm kernel
x86_64 AMD kernel optimized for performance & hypervisor usage
Stars: ✭ 32 (-64.84%)
Mutual labels: svm
Rumble
⛈️ Rumble 1.11.0 "Banyan Tree"🌳 for Apache Spark | Run queries on your large-scale, messy JSON-like data (JSON, text, CSV, Parquet, ROOT, AVRO, SVM...) | No install required (just a jar to download) | Declarative Machine Learning and more
Stars: ✭ 58 (-36.26%)
Mutual labels: svm
Vehicle counting hog svm
Vehicle detection, tracking and counting by SVM is trained with HOG features using OpenCV on c++.
Stars: ✭ 82 (-9.89%)
Mutual labels: svm
Online Svr
Implementation of Accurate Online Support Vector Regression in Python.
Stars: ✭ 52 (-42.86%)
Mutual labels: svm
Nlp Journey
Documents, papers and codes related to Natural Language Processing, including Topic Model, Word Embedding, Named Entity Recognition, Text Classificatin, Text Generation, Text Similarity, Machine Translation),etc. All codes are implemented intensorflow 2.0.
Stars: ✭ 1,290 (+1317.58%)
Mutual labels: svm
Vehicle Detection And Tracking
Udacity Self-Driving Car Engineer Nanodegree. Project: Vehicle Detection and Tracking
Stars: ✭ 60 (-34.07%)
Mutual labels: svm
Defect-Detection-Classifier
Defect detection of water-cooled wall. The sample is hard to collect, so we only have a little dataset
which includes 320 training images(160 normal+ 160 defect) and 80 testing images(40 normal+ 40 defect).
The image size is 256*256. The dataset is collected by Dong Jin. Thanks the advice from Yu Fang about
the using of gcForest.
dataset
the above three images are normal examples and the below are defect.
Classifier
We use Support Vector Machine(SVM) with different feature extractors, deep forest and Convolutional Neural
Network to train the classifier.
-
Gauss filter+LBP+SVM(rbf kernel)
Use Gaussian filter and laplacian operator to denoise and extracts edges, then LBP(Local Binary Patt- ern) extract features of preprocessed images as the input of SVM.
-
CNN+SVM(rbf kernel)
Use VGG16 to extract features as the input of SVM., the weight of VGG16 is trained on ImageNet.
-
simple CNN(3 Conv+1 FC)
Build a simple neural network to train. The network consists of three convolutional layers and a fully connected layer.
-
transfer Learning(VGG16)
Use VGG16 to extract features as input of a simple network that consists of a fully-connected layer.
-
Neural Network Search
Use NNS to search a best network.
-
gcForest
Use deep forest(Only cascade forest structure/With multi-grained forests) to train the ensemble classifier.
Result
classifier | accuracy |
---|---|
Gauss filter+LBP+SVM(rbf kernel) | 97.25% |
CNN+SVM(rbf kernel) | 71.25% |
simple CNN(3 Conv+1 FC) | 72.50% |
transfer Learning(VGG16) | 81.25% |
Neural Network Search | 82.28% |
gcForest (without multi-grained forests) | 80.00% |
gcForest (with multi-grained forests, i=8) | 88.75% |
run
Dependencies
-
AutoKeras
Currently, Auto-Keras is only compatible with: Python 3.6. And we need to install the depedencies under python3. [2019.5.10] -
others
pip install -r requirements.txt
run
# read README.md in models folder and download weight file of pre-trained VGG on the ImageNet dataset.
# dataset
cp -rf normal_add/* ./normal
rm -rf normal_add/
cp -rf defect_add/* ./defect
rm -rf defect_add
# CNN+SVM(rbf kernel)
python cnnSVM.py
# simple CNN(3 Conv+1 FC)
python CNNclassifier.py
# transfer Learning(VGG16)
python transferLearning.py
# gcForest (without multi-grained forests)
python ./data/train/write_label.py
python ./data/test/write_label.py
python ./gcForest/demo_Defect-Detection-Classifier.py --model ./gcForest/demo_Defect-Detection-Classifier-ca.json
# gcForest (with multi-grained forests, i=8)
python ./gcForest/demo_Defect-Detection-Classifier.py --model ./gcForest/demo_Defect-Detection-Classifier-gc8.json
# Neural Network Search, copy autokeras dir to current path after it is installed from source.
python ./data/train/write_label2.py
python ./data/test/write_label2.py
python3 autoCNNclassifier.py
reference
scikit-learn tutorial Building powerful image classification models using very little data
Note that the project description data, including the texts, logos, images, and/or trademarks,
for each open source project belongs to its rightful owner.
If you wish to add or remove any projects, please contact us at [email protected].