# infinte-aditya/gastrovision-4.0 architecture

> Architecture diagram of the GitHub repository infinte-aditya/gastrovision-4.0, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-11-17). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/infinte-aditya/gastrovision-4.0
- Repository: https://github.com/infinte-aditya/gastrovision-4.0

## Overview

GastroVision is a local Windows desktop application and research prototype for analyzing endoscopy videos. A user selects a video in the PyQt interface; the local API accepts it and invokes the analysis pipeline, which extracts and filters frames, classifies them, computes a temporal verdict, and generates explanatory heatmaps for selected frames. The interface retrieves and presents the results. The README also describes model training and shared preprocessing as a research capability. Edges follow excerpted calls or documented behavior; training links are limited to README evidence.

Main entry points: [`backend/main.py`](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/main.py) (Analysis pipeline), [`backend/paths.py`](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/paths.py) (Analysis outputs), [`frontend/gastrovision_single.py`](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/frontend/gastrovision_single.py) (Desktop UI), [`backend/server.py`](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/server.py) (Local API)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_desktop["Desktop application"]
  node_ui["Desktop UI"]
  node_appinfo["App information<br/>[app_info.py]"]
  node_api["Local API<br/>[server.py]"]
end

subgraph group_analysis["Video analysis"]
  node_pipeline["Analysis pipeline<br/>[main.py]"]
  node_extract["Frame extraction<br/>[frame_extractor.py]"]
  node_select["Frame quality filter<br/>[frame_selector.py]"]
  node_frame_score["Frame scoring<br/>[frame_scoring.py]"]
  node_video_score["Video verdict<br/>[video_scoring.py]"]
end

subgraph group_inference["Inference and explanation"]
  node_predictor["ONNX predictor<br/>[predictor.py]"]
  node_preprocess["Image preprocessing<br/>[preprocess.py]"]
  node_heatmap["CAM heatmaps<br/>[heatmap.py]"]
end

subgraph group_storage["Local data"]
  node_paths["Data paths<br/>[paths.py]"]
  node_outputs["Analysis outputs<br/>[paths.py]"]
end

subgraph group_research["Model research"]
  node_training["Model training"]
  node_ml_preprocess["Training preprocessing"]
end

node_user(("Video analyst"))
node_ffmpeg["FFmpeg"]

node_user -->|"selects video"| node_ui
node_ui -->|"reads info"| node_appinfo
node_ui -->|"sends requests"| node_api
node_api -->|"invokes analysis"| node_pipeline
node_pipeline -->|"extracts frames"| node_extract
node_pipeline -->|"filters frames"| node_select
node_pipeline -->|"classifies frames"| node_predictor
node_pipeline -->|"summarizes video"| node_video_score
node_pipeline -->|"computes stats"| node_frame_score
node_pipeline -->|"generates heatmaps"| node_heatmap
node_predictor -->|"preprocesses images"| node_preprocess
node_heatmap -->|"requests CAM"| node_predictor
node_extract -->|"decodes video"| node_ffmpeg
node_pipeline -->|"resolves data paths"| node_paths
node_api -->|"stores uploads"| node_paths
node_paths -->|"provides output location"| node_outputs
node_video_score -->|"reads frame scores"| node_outputs
node_frame_score -->|"reads and writes scores"| node_outputs
node_heatmap -->|"writes heatmaps"| node_outputs
node_training -->|"uses preprocessing"| node_ml_preprocess
node_ml_preprocess -->|"shares implementation"| node_preprocess

click node_ui "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/frontend/gastrovision_single.py"
click node_appinfo "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/frontend/app_info.py"
click node_api "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/server.py"
click node_pipeline "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/main.py"
click node_extract "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/video_processing/frame_extractor.py"
click node_select "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/video_processing/frame_selector.py"
click node_frame_score "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/scoring/frame_scoring.py"
click node_video_score "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/scoring/video_scoring.py"
click node_predictor "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/model/predictor.py"
click node_preprocess "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/model/preprocess.py"
click node_heatmap "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/visualization/heatmap.py"
click node_paths "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/paths.py"
click node_outputs "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/paths.py"
click node_training "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/ml/gastrovision_train.ipynb"
click node_ml_preprocess "https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/ml/gastrovision_preprocess.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_ui,node_appinfo,node_api toneBlue
class node_pipeline,node_extract,node_select,node_frame_score,node_video_score toneAmber
class node_predictor,node_preprocess,node_heatmap toneMint
class node_paths,node_outputs toneRose
class node_training,node_ml_preprocess,node_user,node_ffmpeg toneIndigo
```

## Components

### Desktop application

- [Desktop UI](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/frontend/gastrovision_single.py): `frontend/gastrovision_single.py`
- [App information](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/frontend/app_info.py): `frontend/app_info.py`
- [Local API](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/server.py): `backend/server.py`

### Video analysis

- [Analysis pipeline](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/main.py): `backend/main.py`
- [Frame extraction](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/video_processing/frame_extractor.py): `backend/video_processing/frame_extractor.py`
- [Frame quality filter](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/video_processing/frame_selector.py): `backend/video_processing/frame_selector.py`
- [Frame scoring](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/scoring/frame_scoring.py): `backend/scoring/frame_scoring.py`
- [Video verdict](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/scoring/video_scoring.py): `backend/scoring/video_scoring.py`

### Inference and explanation

- [ONNX predictor](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/model/predictor.py): `backend/model/predictor.py`
- [Image preprocessing](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/model/preprocess.py): `backend/model/preprocess.py`
- [CAM heatmaps](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/visualization/heatmap.py): `backend/visualization/heatmap.py`

### Local data

- [Data paths](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/paths.py): `backend/paths.py`
- [Analysis outputs](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/backend/paths.py): `backend/paths.py`

### Model research

- [Model training](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/ml/gastrovision_train.ipynb): `ml/gastrovision_train.ipynb`
- [Training preprocessing](https://github.com/infinte-aditya/gastrovision-4.0/blob/dev/ml/gastrovision_preprocess.py): `ml/gastrovision_preprocess.py`

### Other components

- **Video analyst**
- **FFmpeg**

## Connections

- Video analyst → Desktop UI: selects video (evidence: `README.md`)
- Desktop UI → App information: reads info (evidence: `frontend/gastrovision_single.py`)
- Desktop UI → Local API: sends requests (evidence: `frontend/gastrovision_single.py`)
- Local API → Analysis pipeline: invokes analysis (evidence: `backend/server.py`)
- Analysis pipeline → Frame extraction: extracts frames (evidence: `backend/main.py`)
- Analysis pipeline → Frame quality filter: filters frames (evidence: `backend/main.py`)
- Analysis pipeline → ONNX predictor: classifies frames (evidence: `backend/main.py`)
- Analysis pipeline → Video verdict: summarizes video (evidence: `backend/main.py`)
- Analysis pipeline → Frame scoring: computes stats (evidence: `backend/main.py`)
- Analysis pipeline → CAM heatmaps: generates heatmaps (evidence: `backend/main.py`)
- ONNX predictor → Image preprocessing: preprocesses images (evidence: `backend/model/predictor.py`)
- CAM heatmaps → ONNX predictor: requests CAM (evidence: `backend/visualization/heatmap.py`)
- Frame extraction → FFmpeg: decodes video (evidence: `backend/video_processing/frame_extractor.py`)
- Analysis pipeline → Data paths: resolves data paths (evidence: `backend/main.py`)
- Local API → Data paths: stores uploads (evidence: `backend/server.py`)
- Data paths → Analysis outputs: provides output location (evidence: `backend/paths.py`)
- Video verdict → Analysis outputs: reads frame scores (evidence: `backend/scoring/video_scoring.py`)
- Frame scoring → Analysis outputs: reads and writes scores (evidence: `backend/scoring/frame_scoring.py`)
- CAM heatmaps → Analysis outputs: writes heatmaps (evidence: `backend/visualization/heatmap.py`)
- Model training → Training preprocessing: uses preprocessing (evidence: `README.md`)
- Training preprocessing → Image preprocessing: shares implementation (evidence: `README.md`)
