prathamd69/kidney-tumor-detection

How prathamd69/kidney-tumor-detection works

Kidney Tumor Detection provides two related workflows: a reproducible training/evaluation pipeline for binary and four-class CT-image classification, and a web API for image upload and inference. In the API, startup initializes both predictors; requests select screening or detailed diagnosis, with class descriptions added to the latter. Training components construct datasets and models, while optimization exports lightweight TFLite models. Some training and multiclass evaluation relationships are documented but their implementations were not excerpted, so those connections are marked unverified.

Main entry points: model_factory.py Model pipeline factory · app.py FastAPI service · prediction.py Binary predictor · file_utils.py File and config I/O

Components 20

Inference service

FastAPI service app.py
Web interface templates/index.html
Image preprocessing src/prediction/prediction.py
Binary predictor src/prediction/prediction.py
Multiclass predictor src/prediction/prediction.py
Class descriptions class_descriptions.json

Training and evaluation

Data processing src/components/data_processing.py
Dataset streaming src/components/build_datasets.py
Binary training src/binaryclassification/binaryclass_training.py
Multiclass training src/multiclassification/multiclass_training.py
Binary evaluation src/binaryclassification/binarymodel_evaluation.py
Multiclass evaluation src/multiclassification/multimodel_evaluation.py

Model construction

Binary TFLite model litemodels/binarymodel.tflite
Multiclass TFLite model litemodels/multimodel.tflite
Model pipeline factory src/components/model_factory.py
Base model builder src/components/build_basemodel.py
TFLite optimizer src/components/model_optimization.py

Shared utilities

File and config I/O src/utils/file_utils.py
Logging src/utils/logger.py

Other components

Image uploader

Connections 18

Image uploader → FastAPI service: uploads image app.py
FastAPI service → Web interface: serves page app.py
FastAPI service → Binary predictor: dispatches screening app.py
FastAPI service → Multiclass predictor: dispatches diagnosis app.py
FastAPI service → Class descriptions: loads descriptions app.py
Binary predictor → Image preprocessing: preprocesses image src/prediction/prediction.py
Multiclass predictor → Image preprocessing: preprocesses image src/prediction/prediction.py
Binary predictor → Binary TFLite model: loads model src/prediction/prediction.py
Multiclass predictor → Multiclass TFLite model: loads model src/prediction/prediction.py
Model pipeline factory → Dataset streaming: builds datasets src/components/model_factory.py
Model pipeline factory → Base model builder: builds model src/components/model_factory.py
Model pipeline factory → File and config I/O: loads data src/components/model_factory.py
TFLite optimizer → Model pipeline factory: configures model src/components/model_optimization.py
Binary evaluation → Model pipeline factory: requests test components src/binaryclassification/binarymodel_evaluation.py
Binary training → Model pipeline factory: uses training pipeline
Multiclass training → Model pipeline factory: uses training pipeline
Multiclass evaluation → Model pipeline factory: uses evaluation pipeline
Data processing → File and config I/O: uses file utilities

More Python diagrams: eth-siplab/egophi · ezykl/r2r-model · fabian57fabian/prototypical-networks-few-shot-learning · fabric8-analytics/fabric8-analytics-server · frequenz-floss/frequenz-sdk-python

Drawn by GitDiagram on September 30, 2026 from the repository's file tree, README and source files. Read as Markdown ·