Problem
Art movement classification requires visual feature learning and interpretable model behavior across visually diverse styles.
FeaturedMachine LearningComputer Vision
Art Style Classification
Deep learning image classification project comparing CNN and transfer learning approaches with Grad-CAM interpretability.
NDA ProtectedCase link soonDemo soon
Approach
Compared convolutional and transfer learning approaches, then used Grad-CAM to inspect which visual regions informed predictions.
Outcome
Created a model comparison workflow with strong baseline performance and interpretability artifacts for visual review.
Technical Stack
TensorFlowKerasCNNViTGrad-CAMImage Processing
Metrics & Signals
90%+ baseline CNN accuracyModel comparisonVisual interpretability