Malaria Detection Using Machine Learning Project

Malaria Detection Using Machine Learning Thesis Ideas

We help scholars to confirm the thesis topics for Malaria Detection Using Machine Learning project that is in current trend. The latest tools that we use will be explained by our resource team. Our team comes up with a topic that provides clarity about your thesis work. Some of the thesis topics that we have developed are sorted below.

1. Malaria Parasite Detection Using CNN-Based Ensemble Technique on Blood Smear Images

Keywords:

Blood smear images, convolutional neural net-works (CNN), computer-aided diagnosis (CADx), deep learning (DL), machine learning (ML), malaria parasite detection, pre-trained models

            Our proposed method utilized ensemble method to improve the accuracy and performance. We utilized NIH Malaria dataset for experimental calculations. We utilized Convolutional Neural Network (CNN) to fastly diagnose red blood cells with malarial parasite infection from segmented microscopic blood smear images that is beneficial in that area with limited medical personnel. 

2. Detection of Malaria Infection Using Convolutional Neural Networks

Keywords:

Deep learning, image classification, malaria, supervised learning

            Our work proposes a novel CNN method to find the cells that is parasitized or uninfected with malaria to aid health personnel that saves the life of affected people. Our proposed CNN method can compare with other pretrained models like VGG-19, ResNet50, DenseNet121 and InceptionV3. Our result displays that the suggested method outperforms best accuracy rate.

3. Edge-Based Segmentation for Accurate Detection of Malaria Parasites in Microscopic Blood Smear Images: A Novel Approach using FCM and MPP Algorithms

Keywords:

Segmentation, Microscopic images, Edge detection

            Our paper suggests a segmentation method to detect parasite cells of malaria with thin blood smear images by utilizing edge-based segmentation. To adjust lighting we used Gamma equalization and to extract infected erythrocytes FCM soft clustering method can be utilized. To enhance edge-based segmentation we utilized MPP method. Our proposed method gives best accuracy.  Our proposed edge-based segmentation method can correctly segment red blood cells in blood smear images.

4. Explainable AI Based Malaria Detection Using Lightweight CNN

Keywords:

Artificial Intelligence (AI), Explainability, Lightweight CNN, SHAP

            Our study offers a customized lightweight Convolutional Neural Network (CNN) to fastly detect malaria from RBC images with explainability. To decrease the processing time a small number of model parameters can be utilized. In addition we can compare this with the state-of-the-arts (SOTA) model. Our prosed methods perform well in both cases. To explain the decision of the suggested method we utilized XAI tool SHAP. Our Lightweight CNN gives the best performance.

5. Improved Malaria Cells Detection Using Deep Convolutional Neural Network

Keywords:

Parasite classification, Feature Extraction, Image Segmentation

            We offer a deep Convolutional Neural Network (CNN) to find the affected malarial cell. The AI model proposed in our work contains three-layered CNN and two-layered dense neural network. The model can seize both minor and important characters by using CNN by retrieving large amount of information from input data. The model can be trained and estimated using binary cross entropy loss function and accuracy metric to access the performance.

6. On Improving Malaria Parasite Detection from Microscopic Images: A Comparative Analytics of Hybrid Deep Learning Models

Keywords:

Classification, CNN-RNN, Image recognition

            In this paper, we develop three hybrid data-driven models and we merge CNN with LSTM, Bi-directional LSTM (Bi-LSTM) with Gated Recurrent Unit (GRU). CNN can be utilized in all three suggested methods to retrieve the related features that are passed to two cascaded layers of RNN in every model can perform as a classifier. CNN-GRU-GRU hybrid model perform better than other models in terms of accuracy. The CNN-LSTM-LSTM was attributed to a low computing.  

7. Malaria Parasite Detection and Classification using CNN and YOLOv5 Architectures

Keywords:

malaria parasite, object detection, classification, YOLOv5

            The aim of our paper is to propose and improve the model to detect malaria parasites. We apply a convolutional Neural Network (CNN) and YOLOv5 methods can detect and classify the malaria with the selected dataset. We utilize a publicly presented dataset consist all images of malaria parasite. After train the CNN model accuracy in detecting affected blood images and the performance will be compared.

8. Detection of Malaria Disease Using Image Processing and Machine Learning

Keywords:

Malaria disease, Blood smear images, Image processing, Computer-aided diagnosis

            Our aim is to produce a computer-aided method to detect the malaria parasite automatically by utilizing image processing and ML methods. Affected or not affected person can be classified by utilizing handcrafted features retrieved from red blood cell images. We have used Adaboost, KNN, DT, RF, SVM and Multinomial Naïve Bayes ML methods on dataset. Adaboost, SVM, RF and Multinomial NB gives best accuracy.

9. Implementation of Malaria Parasite Detection and Species Classification Using Dilated Convolutional Neural Network

Keywords:

Dilated, Parasite, Species

            Our paper offers a Dilated Convolutional Neural Network to detect the malaria parasite and classify the species by utilizing blood smear images. A direct classification can execute three convolutional layers and convolutional 2D for convolution process when the dilation rate of 2 can be utilized for convolutional layers. The model can be trained with publicly available dataset with high performance accuracy.

10. Analysis of different Machine Learning and Deep Learning Techniques for Malaria Parasite Detection

Keywords:

Parasitized, Decision tree, K nearest neighbour, Random Forest, VGG19, Resnet50V2

            The goal of our paper is to compare the machine learning methods namely KNN, DT, LR and RF and execute transfer learning with deep learning methods VGG19, modified Resnet50 to increase the accuracy performed with ML methods that proposes the best model for predicting malaria by only watching the blood cell image.