I am a Master's student in Image Processing and Computer Vision at The University of Electro-Communications, Tokyo, working on deep learning methods for medical image analysis with a focus on reliability and interpretability. I am also reproducing and studying code from previously published research papers to better understand existing methods, verify reported results, and build fairer comparisons for future model development.

Alongside research, I have professional experience building production mobile applications, especially in health-tech systems involving medical image capture, preprocessing, AI pipeline integration, secure patient workflows, and cloud-connected mobile architecture.

• Retinal image analysis and medical image classification
• CNN and hybrid CNN–Vision Transformer models
• Explainable AI using Grad-CAM
• Multi-label classification and evaluation strategies
• Mobile health-tech systems and AI-assisted workflows
• Edge-ready and privacy-aware healthcare applications

• Developing deep learning pipelines for retinal disease diagnosis using public fundus imaging datasets.
• Comparing CNN architectures such as ConvNeXt, EfficientNet, ResNet, DenseNet, Inception, and MobileNet.
• Exploring hybrid CNN–Transformer models for improved performance and robustness.
• Studying evaluation metrics including AUC, mAP, and F1-score, along with threshold tuning strategies.
• Maintaining structured experiment logs and reproducible research workflows.

Built mobile features for secure lesion image capture, patient workflows, dermatologist consultation, and AI-assisted skin analysis. Worked on medical image preprocessing, body-part cropping, selfie-camera scanning, image normalization, Firebase/Google Cloud integration, and Vertex AI-based classification workflows.

Contributed to production mobile application development for a business collaboration platform supporting real-time communication, secure information sharing, smart notifications, and team productivity workflows.

Personal Android project built to explore modern Android development practices including Jetpack Compose, Kotlin, Google Play Console guidelines, Gradle, CI/CD, Detekt, and WorkManager.

Python PyTorch TensorFlow Keras CNNs Vision Transformers Grad-CAM OpenCV scikit-learn Medical Image Classification Multi-label Classification Model Evaluation

NumPy Pandas Matplotlib Jupyter LaTeX Overleaf Experiment Tracking Literature Review Reproducibility Analysis

Kotlin Java TypeScript JavaScript Dart Swift Android Jetpack Compose React Native Flutter REST APIs Firebase Google Cloud Vertex AI

Git GitHub GitLab Bitbucket Jira VS Code PyCharm IntelliJ IDEA Android Studio Figma Adobe XD

• Facial Expression Recognition with PyTorch - Coursera
• Deep Learning with PyTorch: Grad-CAM - Coursera
• Deep Learning with PyTorch: Image Segmentation - Coursera
• Machine Learning - Stanford University / Coursera
• Data Science with Python - IBM
• Creating a Great User Experience for Mobile Apps
• Android Development Certification

• GitHub: github.com/iNuman
• LinkedIn: linkedin.com/in/-inuman
• Email: inuumaan@gmail.com

Building reliable medical AI and production-ready mobile health systems.