AI
Cloud image classification with model comparison
A vision app that classifies uploaded sky images, letting you run the same picture through three different CNN architectures.
Own buildTeam build
01The problem
What it had to solve.
Architecture choice is usually argued from benchmarks run on somebody else’s data. Putting the candidates side by side on your own inputs is a far better way to decide.
02The build
How it was put together.
- Three CNN architectures trained and saved as separate models — AlexNet, LeNet and ResNet
- Flask upload flow with extension validation and safe filename handling
- OpenCV preprocessing — resize to the input size each architecture expects, then normalise
- Model chosen at request time, so one image can be compared across all three
- Per-model predictions exported for offline comparison
Stack
TensorFlowKerasOpenCVFlaskPython
What it demonstrates
- Comparing architectures empirically rather than by reputation
- Image preprocessing handled correctly per model input shape
- A trained model wrapped in something a non-technical person can actually use
Next step
Tell us what you are trying to build.
A first call is thirty minutes, costs nothing, and ends with a straight answer about whether we are the right group for it — plus the names of the specialists who would actually do the work.