ACADEMIC PROJECT · MACHINE LEARNING
Healthier plants.
Smarter decisions.
A computer vision approach to plant disease detection, built with Python, TensorFlow, and OpenCV.
Python / TensorFlow / OpenCV
Read work details
- Built a plant disease classification solution using Python, TensorFlow, OpenCV, Pandas, and NumPy, performing image collection, dataset organization, preprocessing, feature engineering, and supervised model development across multiple plant disease categories.
- Developed and trained convolutional neural network models, tuning hyperparameters and comparing baseline and optimized architectures to improve classification accuracy, robustness, and overall model performance.
- Applied image augmentation techniques including rotation, flipping, and scaling to expand training data diversity, reduce overfitting, and strengthen model generalization across disease categories.
- Built a reproducible preprocessing-to-evaluation pipeline and optimized data-processing workflows to streamline experimentation, improve processing efficiency, validate structured dataset expansion, and accelerate model iteration.
- Evaluated and analyzed model results using accuracy, loss curves, confusion matrices, and misclassified samples, identifying performance gaps and guiding preprocessing and model improvements; documented experimental results and performance comparisons across iterations.