app.py

import streamlit as st

st.set_page_config(
    page_title = "SmartVision AI",
    page_icon  = "🤖",
    layout     = "wide",
)

CLASS_NAMES = [
    "airplane", "bed", "bench", "bicycle", "bird",
    "bottle", "bowl", "bus", "cake", "car",
    "cat", "chair", "couch", "cow", "cup",
    "dog", "elephant", "horse", "motorcycle", "person",
    "pizza", "potted_plant", "stop_sign", "traffic_light", "truck"
]

st.title("🤖 SmartVision AI")
st.subheader("Intelligent Multi-Class Object Recognition System")
st.markdown("---")

col1, col2, col3, col4 = st.columns(4)
with col1:
    st.metric("Total Classes",   "25")
with col2:
    st.metric("Training Images", "2,500")
with col3:
    st.metric("CNN Models",      "4")
with col4:
    st.metric("Detection Model", "YOLOv8")

st.markdown("---")

col1, col2 = st.columns(2)
with col1:
    st.subheader("📋 Project Overview")
    st.markdown("""
    SmartVision AI is a comprehensive computer vision platform that combines:
    - **Transfer Learning** using 4 CNN architectures
    - **Object Detection** using YOLOv8 with bounding boxes
    - **25 Object Classes** from the COCO dataset
    - **Real-time Inference** for immediate predictions
    """)
    st.subheader("🏗️ Tech Stack")
    st.markdown("""
    - **Deep Learning** : TensorFlow / Keras
    - **Detection**     : Ultralytics YOLOv8
    - **Dataset**       : COCO 2017 (25-class subset)
    - **Deployment**    : Streamlit + Hugging Face Spaces
    """)

with col2:
    st.subheader("📊 Model Performance")
    results = {
        "VGG16"         : 64.27,
        "ResNet50"      : 73.87,
        "MobileNetV2"   : 69.60,
        "EfficientNetB0": 72.53,
    }
    for model, acc in results.items():
        st.metric(model, f"{acc:.2f}%")

st.markdown("---")
st.subheader("🎯 25 Object Classes")
categories = {
    "🚗 Vehicles"       : ["car","truck","bus","motorcycle","bicycle","airplane"],
    "👤 Person"         : ["person"],
    "🚦 Outdoor"        : ["traffic_light","stop_sign","bench"],
    "🐾 Animals"        : ["dog","cat","horse","bird","cow","elephant"],
    "🍽️ Kitchen & Food" : ["bottle","cup","bowl","pizza","cake"],
    "🪑 Furniture"      : ["chair","couch","bed","potted_plant"],
}
cols = st.columns(3)
for i, (category, classes) in enumerate(categories.items()):
    with cols[i % 3]:
        st.markdown(f"**{category}**")
        st.markdown(", ".join(classes))

st.markdown("---")
st.subheader("🗺️ Navigation Guide")
col1, col2, col3, col4 = st.columns(4)
with col1:
    st.info("🔍 **Classification**\nAll 4 CNN models")
with col2:
    st.info("📦 **Detection**\nYOLO bounding boxes")
with col3:
    st.info("📊 **Performance**\nModel comparison")
with col4:
    st.info("ℹ️ **About**\nProject details")

st.markdown("---")
st.caption("Built with TensorFlow, YOLOv8 and Streamlit")