Twilight Techy
Ml

Attention-Based Malaria & TB Screening

Five architectures compared properly, with significance testing

TensorFlow
Keras
CBAM
Grad-CAM
OpenCV
scikit-learn
View Code
Attention-Based Malaria & TB Screening

Project Overview

A deep learning framework for automated screening of malaria in blood smears and tuberculosis in chest X-rays, built around Convolutional Block Attention Modules. The interesting part is not any single model but the comparison: five architectures evaluated under matched conditions, with statistical testing, interpretability, and deployment cost all measured rather than assumed.

Client

Research project

Role

Implementation (attention modules, pipeline, evaluation suite)

Completed

July 2026

Duration

Research implementation

Technologies Used

Frontend

Jupyter
Matplotlib
Seaborn

Backend

TensorFlow
Keras
scikit-learn
statsmodels
OpenCV

Deployment

Kaggle GPU
Google Colab

Other Tools

CBAM
Grad-CAM
ResNet50
VGG16
MobileNetV2
DenseNet121

Related Projects

EV Range Estimation
Ml

EV Range Estimation

Range prediction for electric tricycles from a 60-second telemetry window, quantized to INT8 and running on an ESP32 inside a 120 KB tensor arena.

TensorFlow
KerasTuner
TFLite Micro
+3 more
All ProjectsView Portfolio
Twilight Techy

Building digital experiences that make a difference.

GitHubLinkedInX (Twitter)Email

Quick Links

  • Home
  • About
  • Projects
  • Contact

Pages

  • Blog
  • GitHub
  • All Projects
  • Resume

© 2026 Twilight Techy. All rights reserved.