import streamlit as st
from PIL import Image
import torch
from transformers import AutoImageProcessor, AutoModelForImageClassification
import pandas as pd
st.set_page_config(
page_title="CALIX",
page_icon="apple",
layout="centered",
initial_sidebar_state="collapsed"
)
# === MAKE IT A PHONE APP ===
st.markdown("""
Know What You Eat - Instantly. ', unsafe_allow_html=True)
col1, col2 = st.columns(2)
with col1:
st.markdown('', unsafe_allow_html=True)
uploaded = st.file_uploader("Upload food photo", type=["jpg","png","jpeg"])
if uploaded:
st.session_state.uploaded_image = uploaded
st.image(uploaded, width=250)
st.markdown('
', unsafe_allow_html=True)
with col2:
st.markdown('Detects 101 Foods:
Estimate Calories ', unsafe_allow_html=True)
uploaded = st.session_state.get("uploaded_image") or st.file_uploader("Upload photo", type=["jpg","png","jpeg"])
if uploaded:
img = Image.open(uploaded).convert("RGB")
st.image(img, width=300)
with st.spinner("Detecting food..."):
inputs = processor(images=img, return_tensors="pt")
with torch.no_grad():
outputs = model(**inputs)
idx = outputs.logits.argmax(-1).item()
name = food_names[idx]
st.success(f"Detected: {name}")
nut = nutrition_db.get(name, nutrition_db["default"])
st.markdown(f"""
Calories: {nut['cal']} kcal/100g
Carbs: {nut['carb']}
Protein: {nut['prot']}
Fat: {nut['fat']}
""", unsafe_allow_html=True)
if st.button("Add to Daily Log"):
st.session_state.history.append({
"Food": name,
"Calories": nut['cal'],
"Carbs": nut['carb'],
"Protein": nut['prot'],
"Fat": nut['fat']
})
st.success("Added!")
nav_buttons(prev="home", next="food_library")
else:
nav_buttons(prev="home")
footer()
# === FOOD LIBRARY ===
elif page == "food_library":
st.markdown('Food Library (History) ', unsafe_allow_html=True)
if st.session_state.history:
df = pd.DataFrame(st.session_state.history)
st.dataframe(df, use_container_width=True)
total_cal = sum([x for x in df["Calories"] if isinstance(x, (int, float))])
st.info(f"Total Calories Logged: {total_cal} kcal")
else:
st.info("No food logged yet.")
nav_buttons(prev="estimate", next="health_tips")
footer()
# === HEALTH TIPS ===
elif page == "health_tips":
st.markdown('Health Tips ', unsafe_allow_html=True)
tip_idx = st.session_state.get("tip_idx", 0)
st.markdown(f'', unsafe_allow_html=True)
col1, col2, col3 = st.columns([1,1,1])
with col1:
if st.button("Previous"):
st.session_state.tip_idx = (tip_idx - 1) % len(health_tips)
st.rerun()
with col3:
if st.button("Next"):
st.session_state.tip_idx = (tip_idx + 1) % len(health_tips)
st.rerun()
nav_buttons(prev="food_library", next="about")
footer()
# === ABOUT ===
elif page == "about":
st.markdown('About CALIX ', unsafe_allow_html=True)
st.markdown("""
CALIX is an AI food detector that identifies 101 foods from photos.
### Features
- Detects exact names: Pizza, Biryani, Appam, Samosa, Idli, Dosa, etc.
- Local nutrition: Calories, Carbs, Protein, Fat (no internet needed)
- History log: Saves all scans with full details
- 50 Health Tips
- Dark theme + navigation
### AI Model
- eslamxm/vit-base-food101 (Public, 101 classes from Food-101 dataset)
- Dataset: 101,000 images, 101 classes
Built by:
Rudhreshwaran, Shreyas, Tiya Singh, Shubham Prasad, Shubham Raj
AMC CSE-AIML | 2025
""")
nav_buttons(prev="health_tips")
footer()